PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 anew—mum"

INTEGRATING ECOLOGY AND SOCIOECONOMICS FOR SPECIES
RESTORATION: FEASIBILITY OF A LOUISIANA BLACK BEAR
REINTRODUCTION IN AND AROUND BIG THICKET NATIONAL
PRESERVE, TEXAS

By

Anita T. Morzillo

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Fisheries and Wildlife and
Program in Ecology, Evolutionary Biology and Behavior

2005

ABSTRACT
INTEGRATING ECOLOGY AND SOCIOECONOMICS FOR SPECIES
RESTORATION: FEASIBILITY OF A LOUISIANA BLACK BEAR
REINTRODUCTION IN AND AROUND BIG THICKET NATIONAL
PRESERVE, TEXAS
By
Anita T. Morzillo
To mitigate effects of human activity on wildlife species, there are increasing

efforts to recover locally extirpated species in human-dominated landscapes. Thus, the
ability to integrate ecological and socioeconomic factors is critical to better understand
needs of both humans and wildlife. The overall goal of this dissertation was to develop
an interdisciplinary approach to quantify spatial and temporal dynamics of ecological and
socioeconomic factors that may affect a Louisiana black bear (Ursus americanus
luteolus) reintroduction in and around Big Thicket National Preserve (BTNP) in
southeastern Texas. The local bear population was extirpated during the early 19005, but
a recent increase in the number of bear sightings in the area has prompted creation of a
black bear management plan and a simultaneous interest in recovering bear populations.
BTNP is a potential location to release bears, but BTNP’S small size and disjunct
distribution require the consideration of private timberlands and national forests in bear
recovery planning. To determine the ecological feasibility, remote sensing data were
used to identify potential bear habitat across the 2.6 million ha study area.
Approximately 1.3 million ha of highly suitable habitat existed on both private and public
lands. At a density of one bear/ 100 ha, and considering core habitat areas, there is

enough habitat for approximately 2,200 bears. Stand management of vegetative species

that are important for bear foraging will ensure adequate food resources for a bear

population. Because humans also inhabit the area, the study of social feasibility is also
necessary. We surveyed 1,006 local residents to gain insight into attitudes toward black
bears and a preferred recovery strategy. Males, younger residents, those with more
knowledge about bears, and participants in passive-appreciative activities related to
wildlife were more likely to have positive attitudes toward bears and to support
increasing the bear population. Concern about the problems that bears may cause was a
significant predictor of a respondent’s preferred management strategy. Two significant
clusters of survey respondents expressed support for a bear recovery: one in proximity to
Angelina National Forest and the other in Orange County. A simple systems model was
developed to integrate ecological (land cover) and socioeconomic (social survey) data.
Model results suggested that national forests might be better potential release sites for
bears than BTNP. An earlier trend of declining timberland area is likely to continue, and
to be the main driver of degradation in bear habitat over time. Consequently, timberland
loss will likely promote urban development. Furthermore, highest ecological feasibility
did not overlap. Ecological feasibility for bear recovery was greatest in Sabine, Trinity,
and San Augustine Counties. However, social feasibility was greatest in Jasper, Tyler,
and Newton Counties. Public outreach aimed at increasing residents’ knowledge about
bears, as well as gaining insight into why local residents do or do not support bear
recovery, are necessary before deciding whether and where bears should be released.
The model framework provides a useful approach for assessing feasibility of a

reintroduction and may be adapted for conservation programs in other locations.

Copyright by
Anita T. Morzillo

2005

ACKNOWLEDGEMENTS

I am extremely fortunate to have had such a positive doctoral research experience
that has brought me into contact with so many great people. First and foremost, I am
extremely grateful to my advisor Jack Liu. Jack enthusiastically supported my desire to
pursue a very non-traditional and ambitious project, and set a high standard for which I
consistently had (and will continue) to challenge myself. I have benefited immensely
both professionally and personally form Jack’s mentorship and will miss him greatly. I
am also extremely grateful to my committee members Kay Holekamp, Brian Maurer,
Angela Mertig, and Shawn Riley for encouragement, support, and advice. My doctoral
process would not have been as successful without them, and I am very fortunate to have
had their guidance for my research and professional development. I am grateful to
Michigan State University, the National Aeronautical Space Administration (NASA)
Earth Systems Science Fellowship Program, Canon-National Park Service Fellowship
Program, National Fish and Wildlife Foundation, Western National Parks Association,
Texas Parks and Wildlife Department, and Safari Club International—Deep Pineywoods
Chapter for providing funding to make this research possible. In addition, I thank the US
Fish and Wildlife Service, National Park Service, and Big Thicket Association for
providing in-kind contributions.

The assistance of a number of individuals was critical to implementation and
completion of this project. I very sincerely thank Kim Borland, Jayson Egeler, Monica
Glysson, Stephaney Keroson, Sarah Panken, Nathan Pfost, Heidi Wade, and Eunice Yu
for their time and efforts in the field and lab with many tedious and not-always-fun data

collection and entry-related tasks. Technical assistance, guidance, and additional support

from the FW Support Staff, Nathan Garner (Texas Parks and Wildlife Department), Bob
Goodwin, Clinton Jenkins, Maxine Johnson (Big Thicket Association), Matthew
Nicholson, Ken Poff, Dennis Propst, and Bill Taylor are also greatly appreciated. Thank
you (“Xie Xie”) to the Liu lab and my friends at MSU for all of the laughs and
friendships.

Finally, to Eric White, thank you so much for all of the love, support, and
encouragement. Going through the graduate school process together has made the
experience so much more rewarding for so many reasons. I look forward to the

continued sharing of life’s adventures with you.

vi

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. x
LIST OF FIGURES .......................................................................................................... xii
CHAPTER 1
BACKGROUND AND RESEARCH TOPICS .................................................................. 1
1.1. Introduction ......................................................................................................... 2
1.2. Impetus for a Louisiana black bear reintroduction in southeastern Texas .......... 6
1.3. Big Thicket National Preserve — an area targeted for bear recovery .................. 8
1.4. Research topics ................................................................................................. 11
1.5. Conceptual framework ...................................................................................... 20
CHAPTER 2
SCALE, CONTEXT, AND HABITAT ASSESSMENT FOR ECOLOGICAL
FEASIBILITY OF A SPECIES REINTRODUCTION ................................................... 25
Abstract ......................................................................................................................... 26
2.1. Introduction ....................................................................................................... 27
2.2. Methods ............................................................................................................. 30
2.2.1. Background ............................................................................................ 30
2.2.2. Study area .............................................................................................. 32
2.2.3. Habitat classification .............................................................................. 32
2.2.4. Habitat analysis ...................................................................................... 35
2.2.5. Habitat connectivity ............................................................................... 37
2.2.6. Core area analysis ....................................................... 38
2.2.7. Estimation of potential black bear population size ................................ 39
2.3. Results ............................................................................................................... 39
2.3.1. Estimation of total potential habitat area inside and outside the preserve
.......................................................................................................................... 39
2.3.2. Estimation of habitat area based on scale of analysis ............................ 40
2.3.3. Variation in patch quantity and size across scales ................................. 41
2.3.4. Habitat connectivity ............................................................................... 42
2.3.5. Distribution of core habitat area ............................................................ 42
2.3.6. Estimation of potential black bear population size ................................ 43
2.4. Discussion ......................................................................................................... 43
CHAPTER 3
IMPORTANCE OF PRIVATE LAND CONSERVATION FOR A BLACK BEAR
RECOVERY ..................................................................................................................... 63
Abstract ......................................................................................................................... 64
3.1. Introduction ....................................................................................................... 65
3.2. Methodology ..................................................................................................... 68
3.2.1. Study area .............................................................................................. 68
3.2.2. Imagery classification ............................................................................ 68
3.3. Results ............................................................................................................... 71

vii

3.3.1. Changes in area and distribution of bear habitat and timberlands over

time ................................................................................................................... 71
3.3.2. Relationship between timberlands and highly suitable habitat over time
.......................................................................................................................... 72
3.3.3. Relationship between timberlands and changes in habitat quality ........ 73
3.3.4. Changes in habitat area and quality among public lands ....................... 73
3.4. Discussion and conclusions .............................................................................. 74
CHAPTER 4
ATTITUDES AFFECTING SUPPORT FOR A PROPOSED BLACK BEAR
REINTRODUCTION IN EAST TEXAS ......................................................................... 89
Abstract ......................................................................................................................... 90
4.1. Introduction ....................................................................................................... 91
4.1.1. Return of the Louisiana black bear ........................................................ 93
4.2. Methods and measurements ............................................................................... 95
4.2.1. Study area ............................................................................................... 95
4.2.2. Survey design .......................................................................................... 95
4.2.3. Dependent variables ................................................................................ 96
4.2.4. Independent variables ............................................................................. 98
4.2.5. Non-response follow-up ....................................................................... 100
4.2.6. Statistical analysis ................................................................................. 100
4.3. Results ............................................................................................................. 101
4.4. Discussion ....................................................................................................... 103
4.5. Conclusion and implications ........................................................................... 111
CHAPTER 5
SOCIOECONOMIC FACTORS AFFECTING SUPPORT FOR BLACK BEAR
RECOVERY MANAGEMENT STRATEGIES ............................................................ 124
Abstract ....................................................................................................................... 125
5.1. Introduction ..................................................................................................... 126
5.2. Methods ........................................................................................................... 128
5.2.1. Study area and survey design ............................................................... 128
5.2.2. Dependent variables ............................................................................. 128
5.2.3. Independent variables .......................................................................... 130
5.2.4. Non-response follow-up ...................................................................... 130
5.2.5. Statistical analysis ................................................................................. 130
5.3. Results ............................................................................................................. 131
5.4. Discussion ....................................................................................................... 135
5.5. Management implications ............................................................................... 144
CHAPTER 6
SPATIAL ANALYSIS OF ATTITUDES TOWARD PROPOSED MANAGEMENT
STRATEGIES FOR A WILDLIFE REINTRODUCTION ............................................ 150
Abstract ....................................................................................................................... 151
6.1. Introduction ..................................................................................................... 152
6.2. Methodology ................................................................................................... 156

viii

6.2.1. Study area and survey design ............................................................... 156

6.2.2. Dependent variables ............................................................................. 156

6.2.3. Non-response follow-up ...................................................................... 156

6.2.4. Spatial analysis .................................................................................... 156

6.2.5. Distance from the preserve .................................................................. 158

6.3. Results ............................................................................................................. 159

6.4. Discussion ....................................................................................................... 160

CHAPTER 7

INTEGRATING ECOLOGY AND SOCIOECONOMICS FOR SPECIES

REINTRODUCTIONS ................................................................................................... 170

Abstract ....................................................................................................................... 171

7.1. Introduction ..................................................................................................... 172

7.2. Methods ........................................................................................................... 176

7.2.1. Study area ............................................................................................ 176

7.2.2. Model design ........................................................................................ 177

7.2.3. Model structure and quantification ...................................................... 178

7.2.4. Simulation scenarios ............................................................................ 189

7.3. Results ............................................................................................................. 190

7.4. Discussion and conclusions ............................................................................ 194

SUMMARY .................................................................................................................... 223

APPENDIX ..................................................................................................................... 229

A. 1. SURVEY PRENOTICE ................................................................................ 230

A. 3. SURVEY INSTRUMENT ............................................................................ 231

A. 3. SURVEY INSTRUMENT ............................................................................ 232

A. 4. REMINDER POSTCARD ............................................................................ 243

A. 5. SURVEY MAILING #2 COVERLETTER .................................................. 243

A. 5. SURVEY MAILING #2 COVERLETTER .................................................. 244

A. 6. NONRESPONSE FOLLOW-UP SURVEY ................................................. 245

LITERATURE CITED ................................................................................................... 247

ix

LIST OF TABLES

Table 2.1. Area and percent of total area of the eight land cover categories derived from
classification of remotely sensed imagery (30-m scale) for estimation of black bear
habitat in and around Big Thicket National Preserve, Texas .................................... 52

Table 2.2. Area of unsuitable, marginal, suitable, and highly suitable bear habitat across
the whole study area, within Big Thicket National Preserve (BTNP), and outside of
BTNP at the scale of imagery classification (30 m). ................................................ 53

Table 2.3. Area (and percent change) of land cover classes and corresponding black bear
habitat at 30-m (0.09-ha) and 1 ha scales. ................................................................ 54

Table 2.4. Mean (iSD) patch size of black bear habitat at 30-m (0.09 ha) and 1 ha scales.
................................................................................................................................... 55

Table 2.5. Spearman rank correlation values (r,)3 for nearest neighbor analysis of
suitable and highly suitable black bear habitat connectivity at 30m and 1 ha scales
within Big Thicket National Preserve (BTNP) and across the whole study area. 56

Table 3.1. Area of land (ha; percent total area) that has remained consistently within a
single habitat class since establishment of BTNP, and area of land (ha; percent of
total area) that was affected by divestment of land by Louisiana—Pacific (LP). ....... 78

Table 3.2. Area of land (ha; percent total area) that has improved or degraded in habitat
quality since establishment of BTNP, and area of land (ha; percent of total area) that
was affected by divestment of land by Louisiana-Pacific (LP). ............................... 79

Table 4.1. Results of factor analysis (principal components analysis with varimax
rotation) used to define dependent variables3 for analysis of attitudes toward black
bears (N = 982) in East Texas. ................................................................................ 114

Table 4.2. Independent variables used and descriptive results (percentages and mean i
SD)a for analysis of attitudes toward black bears, wildlife managers, and an increase
in the black bear population in East Texas. ............................................................ l 15

Table 4.4. Descriptive results (Percentages and mean i SD) of dependent variablesa
(weighted to account for oversampling of rural residents) used to assess attitudes
toward a black bear reintroduction in southeastern Texas. ..................................... 118

Table 4.5. Results for bivariate statistical analysesa for variables used in southeastern
Texas black bear attitudes survey. .......................................................................... 120

Table 4.6. Multivariate analysis predicting ordinary least squares regression modela for
attitudes toward black bears , wildlife managersc, and increasing the black bear
population sized. ...................................................................................................... 121

Table 5.1. Independent variables used and descriptive resultsa for analysis of attitudes
toward black bears, wildlife managers, and an increase in the black bear population
in East Texas. .......................................................................................................... 146

Table 5.2. Descriptive results of explanatory dependent variables (weighted to account
for oversampling of rural residents) used to designate particular black bear recovery
strategies in East Texas. .......................................................................................... 147

Table 5.3. Results for bivariate statistical analyses8 for variables used for assessing
support for black bear management strategies. ....................................................... 148

Table 5.4. Multivariate analysis predicting ordinary least squares regression model3 for
assessing support for black bear management strategies. ....................................... 149

Table 6.1. Analysis of attitudes toward potential black bear recovery management
strategies using the spatial scan statistic. ................................................................ 164

Table 7.1. Statistically significant survey variables for predicting attitudes toward black
bears that were used for construction of the bear model’s social component ......... 200

Table 7.2. Actual median age (from census data), verified median age (from simulation
of present) and estimated (simulated) changes in median age of residents within
counties of the study area for years 2010 and 2020. ............................................... 201

Table 7.3. Actual percent males (from census data), verified percent males (from
simulation to present) and estimated (simulated) changes in percent males of
resident populations within counties of the study area for years 2010 and 2020. .. 202

Table 7.4. Present and projected (years 2010 and 2020) ecological rankings of counties
for bear recovery based on two sets of 100 simulations and multi-criteria analysis.
................................................................................................................................. 203

Table 7.5. Present and projected (years 2010 and 2020) social rankings of counties for
bear recovery. .......................................................................................................... 205

xi

LIST OF FIGURES

Figure 1.1. Spatial distribution of Big Thicket National Preserve’s administrative units.
................................................................................................................................... 23

Figure 1.2. Conceptual framework for determining the feasibility of a Louisiana black
bear reintroduction in and around Big Thicket National Preserve, Texas. Solid
arrows indicate relationships considered in this research. The dashed arrow
indicates a relationship that will likely be more pronounced if and when bears are
reintroduced. ............................................................................................................. 24

Figure 2.1. The 12-county region in southeastern Texas used to evaluate land cover and
potential habitat for a black bear reintroduction. These counties include and
surround Big Thicket National Preserve (shaded areas in oval). .............................. 57

Figure 2.2. Distribution (shown in black) of (a) unsuitable, (b) marginal, (c) suitable, and
(d) highly suitable black bear habitat throughout the study area in southeastern
Texas. Large water bodies are shaded in light gray. ................................................ 59

Figure 2.3. Distribution (log frequency) of (a) unsuitable, (b) marginal, (c) suitable, and
(d) highly suitable patches of black bear habitat by area (ha). Confidence intervals
were defined at the 95% level. .................................................................................. 60

Figure 2.4. Distribution of core habitat area at (a) 1 ha and (b) 30 m scales, designated by
dark gray shaded areas. Note that no core habitat exists at the 30 m scale, even
within BTNP (outline). ............................................................................................. 62

Figure 3.1. (a) The 12-county study area in southeastern Texas. The southern portion of
Liberty County was excluded from analysis because of absence of timberlands and
almost complete unsuitable habitat as a result of extensive agricultural and
residential land use. (b) USDA proclamation boundaries for the Sam Houston, Davy
Crockett, Angelina, and Sabine National Forests (striped; clockwise from bottom
left) and Big Thicket National Preserve (dark shaded). ............................................ 80

Figure 3.2. Area (1,000 ha) of highly suitable, suitable, marginal, and unsuitable bear
habitat at lO-year intervals since establishment of BTNP. ....................................... 82

Figure 3.3. Area (1 ,000 ha) of land within each habitat class across the study area
compared to land contained within timberlands since establishment of BTNP. ...... 83

Figure 3.4. Changes in distribution of timberlands (red) across the study area, since
establishment of BTNP, based on available timberland ownership data. Also
illustrated are public lands including Big Thicket National Preserve (blue) and
USDA National Forest proclamation boundaries (green). (a) During the late 19603,
before establishment of BTNP, timberlands were concentrated within the
northwestern portion of the study area. (b) During the 19905, timberlands became

xii

more evenly distributed across the study area. (0) Orange areas illustrate land
divested by Louisiana-Pacific in 2002, which are largely concentrated in the central
and eastern portions of the study area. ((1) Remaining timberland in 2002 after
divestment of lands owned by Louisiana-Pacific. County boundaries are designated
by black lines. ........................................................................................................... 84

Figure 3.5. Area (1,000 ha) of land divested by Louisiana Pacific, Inc., in 2002. Among
timberlands, divestment affects approximately 20-25% of land within each habitat
class. .......................................................................................................................... 87

Figure 3.6. Area (1,000 ha) of public land within each habitat class across the study area
since establishment of BTNP. ................................................................................... 88

Figure 4.1. The 12—county region in southeastern Texas used to identify attitudes toward
black bears, wildlife managers, and increasing the black bear population. Survey
participants were selected from these counties. ...................................................... 123

Figure 6.1. The 12-county region in southeastern Texas used to assess the spatial
distribution of resident attitudes toward potential black bear recovery strategies from
which survey participants were selected. These counties include and surround Big
Thicket National Preserve (shaded areas in circle). ................................................ 167

Figure 6.2. Statistically significant clusters of non-support responses for the Natural
(solid squares) and the strongly disagree responses for N0 bear (solid triangles)
strategies for black bear recovery. Also shown are the USDA proclamation
boundaries for the (clockwise from bottom left) Sam Houston, Davy Crockett,
Angelina, and Sabine National Forests (striped) and Big Thicket National Preserve
(shaded). .................................................................................................................. 169

Figure 7.1. The 12 counties (outline, labeled) in southeastern Texas included within the
study area, as well as BTNP (solid shaded) and proclamation boundaries of four
National Forests (hatched). ..................................................................................... 206

Figure 7.2. Conceptual framework for ranking counties within the study area as
ecologically and socially feasible for a black bear reintroduction. Solid arrows
indicate the flow of data through the framework. ................................................... 207

Figure 7.3. Histogram illustrating normally distributed function for rates of timberland
gains and loses across counties of the study area. .................................................. 208

Figure 7.4. Example model structure of the timberland component. “Angelina timber
current” indicated the area of timberland within Angelina County at time t.
“Angelina rate of timberland gain or loss” acted as criteria to designate whether
timberland was gained or lost at each time interval as determined by random number
generation. “Angelina change in timberland area between t and t + 1” quantified the
total amount of timberland gained or lost in Angelina County for each time interval.

xiii

The total amount of timberland gained or lost in time t was then added to “Angelina
timberland current” to calculate “Angelina timberland t + 1.” This model structure
was repeated for each county within the study area ................................................ 209

Figure 7.5. Example model structure of the habitat component. “Angelina habitat
current” indicated the area of total habitat within Angelina County at time 1.
“Angelina change in habitat area between t + 1” quantified the total amount of
habitat gained or lost in Angelina County for each time interval. The total amount
of habitat gained or lost in time t was then added to “Angelina habitat current” to
calculate “Angelina habitat t + 1.” This model structure was repeated for each
county within the study area. .................................................................................. 210

Figure 7.6. Example model structure of the social component. “Angelina social current”
indicated the current value of age and gender social variables used in the model at
time 1. “Angelina change in social variable” acted as criteria to designate change in
the social variable at each time interval as determined by random number
generation. “Angelina change in social variable area for t and t + 1” quantified the
total change in each social variable in Angelina County for each time interval. The
total amount of change in each social variable in time t was then added to “Angelina
social current” to calculate “Angelina social t + I .” This model structure was
repeated for each county within the study area. ...................................................... 21 1

Figure 7.7. Observed versus predicted (:tSE) estimates of timberland area across all
counties of the study area ........................................................................................ 212

Figure 7.8. Observed versus predicted (iSE) estimates of highly suitable habitat area
across all counties of the study area. ....................................................................... 213

Figure 7.9. Observed versus predicted (iSE) estimates of suitable habitat area across all
counties of the study area. ....................................................................................... 214

Figure 7.10. Observed versus predicted (iSE) estimates of marginal habitat area across
all counties of the study area. .................................................................................. 215

Figure 7.1 1. Observed versus predicted (iSE) estimates of unsuitable habitat area across
all counties of the study area. .................................................................................. 216

Figure 7.12. Present (2000) and projected (mean 3: SE) timberland area across counties
of the study area in 2010 and 2020. ........................................................................ 217

Figure 7.13. Present (2000) and projected (mean i SE) highly suitable habitat area across
counties of the study area in 2010 and 2020. .......................................................... 218

Figure 7.14. Present (2000) and projected (mean :t SE) suitable habitat area across
counties of the study area in 2010 and 2020. .......................................................... 219

xiv

Figure 7.15. Present (2000) and projected (mean t SE) marginal habitat area across
counties of the study area in 2010 and 2020. .......................................................... 220

Figure 7.16. Present (2000) and projected (mean i SE) unsuitable habitat area across
counties of the study area in 2010 and 2020. .......................................................... 221

Figure 7.17. Rank sum results for quantifying the difference in county rankings for
ecological and social feasibility of a black bear reintroduction. A score of 1-4 on
either the ecological or social feasibility resulted in a rank of “high,” 5-8 was a
“medium,” and 9-12 was a “low.” Area of highly suitable habitat within timberland
for each county is shown in parentheses. ................................................................ 222

XV

CHAPTER 1

BACKGROUND AND RESEARCH TOPICS

1.1. Introduction

There is an increasing recognition that large carnivores are valuable members of
the ecological community (Sillero-Zubiri and Laurenson 2001). While the area of habitat
for many species is decreasing, efforts are increasing to recover populations of locally
extirpated carnivore species (Reading and Clark 1996). As human development rapidly
expands into previously rural areas, protected areas (e. g., nature reserves) and other
public lands are often identified as among the last available habitats for species recovery
initiatives (Woodroffe 2001a). Unfortunately, recovery is often more difficult for large
species that require contiguous areas of habitat larger in size than within a species
geographical range (Pelton 1986; Meffe and Carroll 1997; Woodroffe 2001a).

Sacredness of reserves is a result of the perception of them as exemplifying
models for both biodiversity conservation and wilderness protection (Pimm and Lawton
1998; Soule and Sanjayan 1998). However, the ability of reserves to protect resources is
not guaranteed. For example, in the United States the National Park Service has
identified categories of internal and external threats (e.g., pollution, visitor physical
impacts such as trail overuse, physical removal of resources such as illegal hunting within
park boundaries, land use such as construction of new visitor services within the reserve,
and road construction that can result in increased vehicle collisions with wildlife) to
reserves and their resources (N PCA 2001). However, there is a severe lack of
quantitative and useful information about many important issues, such as l) where the
threats originate (e. g. result of a specific land use activity), 2) why the threats have
occurred, 3) how these threats have changed over time and across space, 4) how much the

impacts of these threats have affected the reserve (e. g. reduction in population size of

resident species, limitation of dispersal, and increase in dispersal mortality), 5) how to
reduce the threats effectively, and 6) how to predict the future dynamics of these threats
and their ecological impacts.

Recently, significant attention has focused on whether reserves effectively
achieve management objectives such as wildlife conservation (Leader—Williams et al.
1990; Soule 1991; Arcese and Sinclair 1997; Meffe and Carroll 1997; Liu et al. 1999;
Cork et al. 2000; Cuarén 2000; Bruner et al. 2001; NCPA 2001). Areas that are remote
(i.e., relatively safeguarded from human activity) with high biological productivity are of
high value for particular species (Carroll et al. 2001). However, external ecological
disturbance (e.g. fragmentation) can result in reserves becoming “habitat islands”
(Brashares et al. 2001; Meffe and Carroll 1997; Soulé 1991). Habitat isolation, which
mimics fragmented habitat patches (Brown 1971; Turner et al. 2001), are often more
pronounced in smaller reserves (Harcourt et al. 2001) and for larger species (Newmark
1996). In addition, complexity of the surrounding land use and socioeconomic conditions
may heavily influence the survival ability of wildlife that travels beyond reserve
boundaries (e.g. Newmark 1996; Woodroffe and Ginsberg 1998; Rivard et al. 2000;
Harcourt et al. 2001).

Mortality that occurs when wildlife travels beyond reserve boundaries is a
particular concern for large home-ranging terrestrial species (Parks and Harcourt 2002).
In fact, human activity and habitat alteration in proximity to reserves has greatly
increased large mammal extinction rates worldwide (Newmark 1996), and is often more
influential on extinction rates than processes within the reserve (Rivard et al. 2000).

Woodroffe and Ginsberg (1998) indicated that [worldwide] 74% of adult large carnivore

mortality, for wide-ranging, generalist species with historical ranges that are both
fragmented and in proximity to reserves, results from human conflict when animals travel
beyond reserve boundaries. Along with wildlife management within a reserve,
conservation efforts must also focus on private land use activities surrounding the reserve
and local socioeconomics that make target species most vulnerable (Pressey and Tully
1994; Parks and Harcourt 2002; Pressey et a1. 2002).

Lack of integrating ecology and socioeconomics has led to many conservation
failures. For example, a great quantity of ecological data has been collected about the
biology of giant panda (A iluropoda melanoleuca) habitat. Until recently, socioeconomic
studies were nonexistent. As a result, the Chinese government, even when providing free
housing, largely failed to move residents out of Wolong Nature Reserve (established in
1975 for panda protection). This resulted from lack of understanding of two important
social issues: 1) elderly residents were hesitant to relocate and adapt to new places, and 2)
there was no available land in proximity to the free housing that this agriculture-based
community could farm (Liu et al. 2001). Integrating ecology and socioeconomics will
increase the effectiveness of policy makers’ abilities to create feasible and flexible
solutions to meet the needs of both wildlife and humans (Liu et al. 2001 ).

Wildlife management in US national parks historically focused on management of
ungulates, and often at the expense of other species (Wright 1999). Over time, an
ecosystem approach has been adopted for reserve management with a focus on the
ecological needs of species. For example, even though Yellowstone National Park
contains enough area to accommodate home ranges of many individual animals,

successful habitat protection for populations of focal species including the grizzly (Ursus

arctos), wolf (Cam's lupus), and wolverine (Gulo gulo) would require an expansion of the
reserve area by 50% (Noss et al. 2002). As is the case on other continents (N ewmark
1996; Harcourt et al. 2001), smaller US reserves are often located in areas of higher
human density, as compared to large reserves (Parks and Harcourt 2002). Gurd et al.
(2001) suggested that to adequately conserve mammals within eastern North American
reserves, a combination of corridors, buffers, and reserves of at least 2,700 km2 would be
necessary. Because ecosystems are dynamic, management goals and potential courses of
action must also be dynamic and continuously evaluated even in instances of incomplete
knowledge about particular species and their behaviors (Wright 1999).

Large species with negatively perceived (by humans) consequences (e. g., danger
to humans) have the potential to provoke negative feelings in humans (Kellert et al.
1985a, b, c). In the US, such attitudes and opinions are often associated with carnivores,
such as black bear (Ursus americanus) and wolves (Canus lupus). As a result,
biologically (i.e., suitable habitat) and socially suitable environments (i.e., human
acceptance of a species) for species are intertwined, which is a challenge for wildlife
managers. A relatively large number of carnivore studies have been completed in US
reserves, but many have focused on wildlife nuisance activity, such as in Yellowstone
and Great Smoky Mountains National Parks (e. g., Beeman and Pelton 1976; Pelton et al.
1976; Garshelis and Pelton 1980; Stiver 1991; Blanchard and Knight 1995). Very little
information is available that quantitatively supports how human activity affects particular
species in proximity to protected areas, particularly as related to species reintroductions.

In a general sense, my dissertation goal was to integrate ecological and

socioeconomic data to determine how human activities affect the ability of reserves to

protect wildlife resources. I extensively researched particular protected areas within the
US to identify a case study that would allow me to achieve this goal. My inquiries led
me to contact Roy Zipp, former Resource Manager for Big Thicket National Preserve.
and Nathan Garner, Regional Director of the Texas Parks and Wildlife Department.

After discussions with them and additional resource managers in southeast Texas about
recovering locally extirpated black bear populations within eastern Texas, my dissertation
evolved into a study of how threats to a protected area will affect recovery of a locally
extirpated species. The overall goal of this research was to use an interdisciplinary
approach to quantify spatial and temporal dynamics of ecological and socioeconomic
processes that may affect a Louisiana black bear (Ursus americanus luteolus)

reintroduction in and around Big Thicket National Preserve (BTNP), Texas.

1.2. Impetus for a Louisiana black bear reintroduction in southeastern Texas
There are 16 documented subspecies of the American black bear (Ursus
americanus; Hall 1981). The Louisiana black bear (Ursus americanus luteolus) has
several morphological characteristics that make it distinct from other black bear
subspecies. Skull morphology, as well as DNA and tissue analysis suggested enough
morphological and genetic difference between U. a. luteolus and other black bear
subspecies for federal protection of U. a. luteolus as a separate subspecies (Kennedy
1989; Neal 1990). In 1992, the USFWS listed the luteolus subspecies as federally
“threatened” (Neal 1990). Based on similarity of appearance, all black bears within the

luteolus historical range are protected (BBCC 1997).

The historical range of the Louisiana subspecies included Louisiana, eastern
Texas, and the southern half of Mississippi (Pelton 1982; Wooding et al. 1994a; Bowker
and Jacobson 1995; BBCC 1997). Black bears were an important component of the
economic history of southeast Texas. Bear meat and oil were widely used by local
Native Americans and sought by riverboat captains. Settler accounts from the 18305-
18405 described years of successful bear hunting throughout the Big Thicket Region
(Abernathy 1967). During the late 18005, bear numbers quickly declined as a result of
bear hunts organized by famous bear hunters including Ben Lilly (who collected
specimens for the US Biological Museum) and Ben and Bud Hooks. Without game laws,
some individual hunters killed more than 100 bears, some of which weighed more than
180 kg. Bears were almost eliminated from coastal southeastern Texas by 1900.
Accounts of settlers suggested the presence of bears in the southeast Texas through at
least 1906 (Truett and Lay 1984), although the last official reports of native bear were a
female and two cubs, and a single large bear, killed in 1919 and 1928, respectively
(Abemethy 1967). In Mississippi, the last recorded breeding occurred in 1976, although
the exact date of extirpation of either subspecies found in Mississippi, luteolus or
americanus, is unknown. Two remnant breeding populations survived within the
bottomland hardwood forests of Louisiana: one within and surrounding the Tensas River
National Wildlife Refuge and adjacent Big Lake Wildlife Management Area in
northeastern Louisiana, and the other within the Atchafalaya Basin in southeastern
Louisiana (Weaver et al. 1990; Marchinton 1995; Bowker and Jacobson 1995; Garner

1996; BBCC 1997; Eastridge 2000). Several sightings have occurred in Mississippi, but

no confirmed breeding population is present (Wooding et al. 1994a; Stinson and Pace
1995, Shropshire 1996; Stinson 1996).

During the mid 19905, Louisiana landowners and members of bear-related
conservation organizations launched a major public outreach and information campaign
to recover statewide bear populations. The Black Bear Conservation Committee and US
Fish and Wildlife Service created recovery plans, and one objective of each plan was to
restore the Louisiana black bear throughout its historical range (Bowker and Jacobson
1995; BBCC 1997). To date, recovery in Louisiana has been successful (Van Why and
Chamberlain 2003), and feasibility analyses were completed for Mississippi with positive
results (Shropshire 1996; Bowman 1999).

Although there is no known breeding population of bears in eastern Texas, the
number of black bear sightings has increased dramatically during the past decade. These
are likely transient bears from bear populations in Louisiana, Arkansas, or Oklahoma.
The number of increased sightings prompted the creation of a bear management plan for
eastern Texas by the Texas Parks and Wildlife Department. The ultimate goal of the
management plan is to restore habitat for the future reestablishment of black bear as a
viable ecosystem component in eastern Texas (TPWD 2005). Short-term objectives
(over the next 10 years) include public coordination, communication, outreach and

information dissemination, habitat management, and research (TPWD 2005).

1.3. Big Thicket National Preserve — an area targeted for bear recovery

Big Thicket National Preserve (BTNP) is a potential target for bear recovery

(Figure 1.1; images in this dissertation are presented in color). Also known as the
“Biological Crossroads of North America” and “America’s Ark,” eleven ecosystems
within the region have been defined by the National Park Service based on dominant
vegetative, geologic, and hydrologic characteristics (Peacock 1994). Native Americans
originally called the region containing what is now BTNP “The Big Woods,” or “The
Impenetrable Woods,” and few permanent settlements were established within it (Gunter
1971; Gunter 1993). While Texas was under Spanish rule during the 18205, the Spanish
government offered land holdings of 840 acres to settlers of European origin. However,
many settlers deserted the area after becoming fi'ustrated with the dense and swampy
forest. Some of the swampiest land remained unclaimed until only recently (Abernathy
1967; Truett and Lay 1984).

After Texas became part of the US (1821), timber and oil were economic
incentives for further settlement of southeast Texas. The first commercial timber
operations began in the 18505 to meet Eastern demand for lumber after depletion of Great
Lakes forests (Truett and Lay 1984). Large scale oil drilling began in 1901 and soon oil
wells were so numerous that they often sank when the ground collapsed under their
weight (Abernathy 1967; Truett and Lay 1984). Today, timber and oil are still important
components of the local economy. Plans to establish BTNP were originally met by
opposition of local residents who believed that the local economy would suffer, visitor
services would force locals out, and children would be eaten by govemment—sponsored
bears and panthers (Gunter 1993).

In 1974, BTNP was established as the nation’s first National Preserve to protect

the 11 unique ecosystems in the area. Since establishment, BTNP has also been

designated as a UNESCO International Biosphere Reserve, an American Bird
Conservancy Globally Important Bird Area (IBA), and member of the United States Man
and Biosphere Program. The preserve’s 12 land and river corridor management units
together total >39,256 ha. Three remaining units (4,290 ha total) have been authorized,
but actual land has not yet been fully acquired.

Because timber management companies still own a large portion of land
surrounding the preserve, forested areas have been maintained across the landscape and
may provide refuge for bears. Two small ecological feasibility studies for the area have
been completed for individual BTNP units, and some private lands surrounding the
preserve (Garner 1996; Epps 1997). However, neither landscape-level ecological nor
socioeconomic analyses have been completed to determine whether a black bear
reintroduction is feasible. Therefore, I developed six focal areas of analysis: (1) use of
remotely sensed imagery to quantify available bear habitat and assess the distribution of
habitat within the study area to determine the ecological feasibility of a black bear
reintroduction, (2) assessment of the dynamics of timberlands to determine if changes in
timberland ownership affect the quantity of bear habitat, (3) implementation of a mail
survey to determine socioeconomic factors that influence local residents’ attitudes toward
black bears and their reintroduction, (4) use of survey data to determine socioeconomic
factors that influenced residents’ selections of preferred management strategy for black
bears, (5) development of a method to assess the spatial distribution of residents’
selections of preferred management strategy, and (6) development of a method to

integrate ecological and socioeconomic data to determine feasibility of a black bear

10

recovery across the study area based on resident attitudes and land use. A brief review

for these focal areas. with a more detailed review within each chapter, is as follows.

1.4. Research topics

For bears to be reintroduced, habitat must be available. Therefore, the first goal
of this study was to use remotely sensed imagery to quantify available bear habitat and
assess the distribution of habitat within the study area to determine the ecological
feasibility of a black bear reintroduction. To date, habitat suitability indices (HSl's) have
been the most commonly used method for evaluating bear habitat (e. g. Landers et al.
1979, Pelton 1982; Rogers and Allen 1987; van Manen 1991). Although HSI's are useful
tools for quantitatively comparing habitat quality between areas of interest, a major
shortcoming of HSI's is that they provide a relative site-specific quantitative assessment
based on predetermined variables that only allows for basic comparisons between
different locations (USFWS 1981; Schamberger et al. 1982). In addition, generalizing
HSI's to spatial and temporal dynamics of landscapes is extremely field intensive and
difficult (Shamberger et al. 1982). Technological advances including geographic
information systems (GIS) and modeling procedures (e. g. Buckland and Elston 1993)
have allowed for more detailed spatial study of bear habitat (Clark et al. 1993; van Manen
and Pelton 1997; Jones 1998; Larkin et al. 2004), but few studies have assessed how
habitat suitability for black bears changes over time and across space (Schooley et al.
1994)

High quality habitat and large tracts of land with minimal human disturbance are

considered highly influential factors of past successful bear reintroductions (Feis et al.

11

1986; Rogers 1986; Stiver 1991; Comly 1993; Riley et al. 1994; Blanchard and Knight
1995). A viable population, adequate area for home ranges, and large trees or dense
understory for denning are important ecological characteristics for maintaining an
established black bear population (Rogers 1987; Rudis and Tansey 1995; Oli et al. 1997).
More specifically, forested lowlands with waterway corridors and a bottomland species
complex containing oaks (Quercus spp.), hickory (Carya spp.), gum (Liquidambar
styraciflua), and cypress (Taxodium spp.) have historically provided the best bear habitat
in the southeastern US (Shropshire 1996; BBCC 1997) by providing understory cover as
well as fruits and mast for forage (Sauer et al. 1969; Landers et al. 1979; Eagle and
Pelton 1980; Maehr and Brady 1984; Smith 1985; Schullery 1986; Weaver 1992;
Stubblefield 1993; Lariviere 2001). Remote forest tracts (i.e., isolated from human
activity), defined by forest areas >0.8 km from development (Rudis 1986), or contiguous
tracts of >1,000 hectares are also desirable (Hellgren and Vaughan 1989a; Brody and
Pelton 1989; Rudis and Tansey 1995; Larkin et al. 2004). Literature has suggested
minimum areas of 7,580 ha for a population of 50 individual bears, 30,300 ha for 200
bears, and 152,000 ha for 1,000 bears (Hellgren and Vaughan 1989a; Rudis and Tansey
1995)

In southeastern Texas, results from Garner (1996) suggested that habitat
suitability for food, availability of cover, and that human impacts would likely not be a
limiting factor for bears. Lack of large denning trees was not considered critical because
bears in the southeastern US often den on the ground within thick undergrowth or slash

piles (Weaver et al. 1990; Marchinton 1995).

12

Epps (1997) completed a more detailed study of two contiguous BTNP units
(Neches Bottom and Jack Gore Baygall), an area of approximately 5,000 ha surrounded
by large tracts of timberland. Based on fall food availability (Smith 1985) and tree den
sites (Rogers and Allen 1987; van Manen 1991), results suggested a density of 4-54
potential den trees/kmz, a sufficient number for a bear population, and an estimated
carrying capacity of 48-86 bears (Pritchard and Robbins 1990). This estimate (0.5 8-1 .40
bears/kmz) includes the range of other carrying capacity estimates for the southeastern
US (0.18-0.77 bears/kmz; Hamilton 1978; Beecham 1980; Smith 1985; Hellgren and
Vaughan 1989b; van Manen 1991).

Nearby forested areas may help support a core population within the preserve
(Johnson and Pelton 1981; Pelton 1986; Epps 1997). Both Garner (1996) and Epps
( 1997) concluded that adequate habitat exists in southeastern Texas for a bear population.
However, bears are potentially far-wandering (Beeman and Pelton 1976) and the area of
BTNP (>3 7,000 ha) is likely too small for a large black bear population. As a result,
individual bears will likely wander between BTNP’s 12 disjunct units (Figure 1.1).
Consequently, the connectivity of habitat, or how close areas of habitat are to each other,
may be a factor in bear movement between preserve units (Beausoleil 1999; Maehr et al.
2003; Larkin et al. 2004). By using remotely sensed imagery, our objectives within the
first research goal (ecological feasibility) were to: (1) quantify how much habitat is
available within the preserve, and how this estimates of habitat change when the preserve
is considered within the context of the surrounding landscape, (2) estimate changes in
habitat area as a result of changing the scale of analysis, (3) assess how changing the

scale of analysis affects connectivity of habitat, (4) identify and quantify core areas (see

13

Methods) of habitat, and whether their distribution change as scale of analysis changes,
and (5) based on habitat area, estimate black bear population size that may exist within
the study area.

To reiterate, as a result of BTNP’s disjunct distribution bears will likely travel
beyond the preserve’s boundaries in search of food, cover, and mates. Therefore, private
land between the preserve’s units will likely be important as black bear habitat as well. A
majority (~70%) of land in southeast Texas is owned by timber companies and managed
for timber production, which has likely been a major influence in the continuous presence
of forestland over time. The subtropical climate allows overall primary productivity to be
high, and fast growth rates of trees within forested areas provides bear cover and forage.
Results from Rudis (1986) suggested that approximately five million ha of timberland
existed in southeast Texas during the 19805, with a large amount of timberland >1 km
away from the nearest road than in Louisiana. The adaptability of bears to diverse
habitats has allowed for easy integration of timber management to meet bear habitat
needs, particularly through maintenance of land access and habitat that may serve as
corridors (Brody and Stone 1985; Hillman and Yow 1986). In fact, bears often denned
within slash piles created from timber harvest or road construction, and thick understory
vegetation (Hamilton and Marchinton 1977; Hillman and Yow 1986; Weaver et al. 1990;
Wooding and Hardisky 1992; Marchinton 1995). Availability [or lack] of denning sites,
such as large trees, did not appear to be a limiting factor for bears in LA (BBCC 1997).

Since Rudis’s survey two decades ago, only Garner (1996) has studied
relationships between timberland and potential bear habitat in southeastern Texas.

However, the objective of Garner’s (1996) study was to complete HSI's for particular

14

expanses of timberland as related to quantifying black bear habitat quality. No research
has been conducted to determine whether the area and distribution of bear habitat among
timberlands has changed since the research by Rudis (1986). Therefore, our second goal
was to assess relationships between timberlands and potential bear habitat, and specific
objectives included adoption of a landscape perspective to (1) quantify how area and
distribution of potential bear habitat and timberlands have changed since establishment of
BTNP, (2) quantify the relationship between highly suitable habitat and distribution of
timberlands, and how this relationship has changed over time, (3) quantify the
relationship between timberlands and areas of consistent, improved, or degraded habitat
quality over time, (4) compare changes in habitat quantity and quality over time among
private lands with such changes among public lands, and (5) assess implications
regarding changes in timberlands and how they may affect attainment of bear
management goals by the Texas Parks and Wildlife Department.

Changing distribution of timberlands is a socioeconomic phenomenon, and
individual attitudes and opinions are often factors in land use decisions (Rudis and
Tansey 1995; McDonald et al. 2001). Timber companies in southeast Texas are
extremely supportive of bear recovery and actively involved in bear management and
recovery efforts. Even so, there is no guarantee that bears will remain within the
boundaries of public lands or timberlands. As a result, neglecting local residents’
attitudes toward bears could have negative impacts on bear recovery efforts. For
example, a 30-year long recovery effort in Arkansas (1958-19905) resulted in an increase
in the black bear populations by >2,500 bears (Rogers 1973; Smith and Clark 1994).

Although this may seem ecologically successful, poaching has been a constant factor in

15

bear population dynamics since the reintroductions (Rogers 1973; Smith et al. 1990;
Shull 1994). It has been hypothesized that some poaching is directly related to the
Arkansas Game and Fish Commission not seeking public input from the outset as a result
of the perception that the public did not support a bear recovery (Smith and Clark 1994).
During the late 19805 and early 19903, the Arkansas Game and Fish Commission became
involved in public awareness programs for bear conservation, and public involvement in
bear conservation has been increasing over time (Smith and Clark 1994).

Public acceptance of wildlife is a major challenge of wildlife management
(DuBrock et al. 1978; Purdy and Decker 1989; Craven et al. 1992; Burton et a1. 1994;
Koch 1994; Locker and Decker 1995; Lohr et al. 1996; Lauber and Knuth 1996; Riley
and Decker 2000a; Riley et al. 2002). Public participation, perceptions, and acceptance
are exceptionally critical to success of carnivore reintroduction efforts (Duda 1986;
Reading and Clark 1996; Kellert 1985b; Enck and Brown 2000; Clark et al. 2003). Bears
are considered charismatic megafauna, and are among species considered human
favorites (Kellert 1985b, c; Hastings and Hammitt 1986; Sellars 1997). Because public
involvement can be the deciding factor in the success (or failure) of a species recovery,
stakeholder support in proximity to target reintroduction areas is critical to success of the
reintroduction (Reading and Kellert 1993; Reading and Clark 1996; Clark et al. 2003).
Analysis of local stakeholder values and attitudes must be completed before a
reintroduction project is pursued so that outreach and public-relations programs may be
custom made for the species, local culture, and geographic region (Reading and Kellert

1993). It is essential that residents of southeastern Texas tolerate bear presence,

16

regardless of whether they individually want bears in the area. Implementation of a
resident survey allowed me to achieve the third, fourth, and fifth goals of this study.

The third goal was to determine socioeconomic factors that influence local
residents’ attitudes toward black bears. Specific objectives were to: 1) identify public
attitudes toward black bears and a potential reintroduction, 2) determine what factors
influence residents’ attitudes toward recovery of black bear in the area, and 3) compare
this information with attitude studies from the remainder of the Louisiana black bear’s
historical range. Past study has suggested that the general public holds existence and
conservation value for bears, and commonly explored attitude variables include (among
others) sex, age, income, education, and location of residence (Decker et al. 1981; Kellert
1994; Hastings and Hammitt 1986; Shropshire 1996; Bowman 1999; Bowman et al.
2001; Peyton et al. 2001). Even in cases of nuisance activities such as campsite pilfering
(Great Smoky Mountains National Park; Pelton et al. 1976) or crop damage (Jonker et al.
1998; Bowman et al. 2001), presence of bears was viewed as an important ecological
and/or economic component of the environment. Louisiana and Mississippi residents
held generally positive attitudes toward black bears and population recoveries in those
states (Shropshire 1996; Bowman 1999; Bowman et a1. 2001; Bowman et al. 2004; Van
Why and Chamberlain 2003).

Even if local residents had generally positive attitudes about black bears,
popularity alone may not be enough to win public support for a reintroduction (Lohr et al.
1996). For example, in Mississippi, public support for increasing the size of the
Mississippi black bear population decreased when a reintroduction was suggested

(Shropshire 1996). Bowman (1999) suggested that more than 60% of both private and

17

corporate landowners did not support a non-assisted (i.e., bears returning on their own)
black bear recovery in Mississippi. Familiarity with bears was a factor in support for
increasing the black bear population in New York (Decker et a1. 1983), Arkansas (Clark
1991 ), and Michigan (Peyton et al. 2001). Therefore, the fourth goal of this study was to
further explore factors that influenced residents’ selection of preferred general
management strategy for black bears if a recovery occurs.

Because it is likely that a variety of attitudes toward potential management
strategies exist throughout southeast Texas, spatial knowledge about attitudes may allow
managers to identify locations where residents’ preferred management strategies are both
favorable for and threaten a reintroduction. For example, residents who want bears in the
area may be more willing to adapt land use practices that favor black bear (e. g.,
maintenance of thick undergrowth on their property). Likewise, residents who oppose
bear presence may pose risks to a successful reintroduction (e. g., shoot bears on sight,
harass individual bears). In addition, if BTNP is a target for a potential bear release site.
it is imperative to know whether residents near the preserve are supportive of bear
recovery efforts.

Spatial analysis of particular occurrences is common in disciplines such as public
health (Knox 2005), environmental valuation (Brown 2005) and wildlife ecology (Kie et
al. 2002). However, spatial research involving human dimensions of wildlife has mainly
focused on comparative studies (Fox et al. 1996; Merrill et al. 1999; Riley and Decker
2000; Peyton et al. 2001) or predictive statistical models (Bowman et a1. 2004). I was
unaware of any studies that had applied spatial analysis of attitudes to assess distribution

of individual survey responses. Therefore, the fifth goal of this study was to assess the

18

spatial distribution of residents’ selection of preferred management strategy.

Specifically, I sought to (1) ascertain the spatial distribution of residents’ attitudes toward
potential bear recovery strategies, and (2) determine whether distance from BTNP is a
factor in resident attitudes.

Finally, I used the data described for goals #1-5 above to achieve the sixth goal of
this study: to integrate ecological and socioeconomic data to determine how resident
attitudes and changes in land use may affect a black bear reintroduction. Goals #1-5
provide insight into factors that may threaten the success of BTNP to provide refuge to
reintroduced black bears. However, they do not allow for simultaneous evaluation of
interrelationships among these factors within the system of study. The first objective to
achieve this goal was to develop a systems model that incorporated bear habitat, land
ownership, and socioeconomic variables that affect attitudes toward black bears in
southeast Texas. Ideally, locations with a greater amount of habitat will also be more
socially feasible and support a reintroduction. We evaluated our model at two fiiture time
points: (1) 10 years into the future (10 years after model parameterization, or 2010), and
( 2) 20 years into the future (20 years after model parameterization, or 2020). The year
2020 coincides with the end of the 15-year term of the present black bear management
plan (TWPD 2005). Our second objective was, based on results of the first objective, to
rank the counties within the study area in regard to both ecological and social feasibility
for the planned black bear recovery. Comparing feasibility between counties will allow
managers to weigh management options based on ecological and social feasibility. For
example, if a particular county contained a high ecological ranking (e. g., large area of

highly suitable habitat) but a low social ranking (e. g., low scores for questions about bear

l9

knowledge), managers will know that habitat is present, but extensive public outreach
will be necessary to determine if bear reintroduction into the county is truly feasible. Just
the opposite, if a particular county contained a low ecological but high social ranking,
managers may prefer to focus public outreach on resident land use management for

improving black bear habitat.

1.5. Conceptual framework

Although black bears are considered habitat generalists (Lariviere 2001), it is
likely that human land use decisions and activities will be the main factors that affect
bear habitat, even if a bear is reintroduced into a protected area (Figure 1.2). Human-
related factors include both direct changes to habitat through land use (e.g., forestry,
urbanization), and indirect drivers of habitat change (e.g., socioeconomic variables such
as income or education). Important characteristics of black bear habitat include quantity
(i.e., area), quality (i.e., highly suitable versus marginal), and location (i.e., spatial
distribution, connectivity; Pelton 1982; Bowker and Jacobson 1995; BBCC 1997; Liu et
al. 1999; Lariviere 2001). If a reintroduction takes place, changes in bear habitat from
season to season may result in bears directly affecting humans. For example, during a
poor mast year, bears may search for food in trash bins and gardens at a greater rate than
during good mast years. Increased nuisance activity may result in negative human
attitudes toward bears and bear behavior.

Land use, human attitudes, local economic conditions, and human demographics
can all affect (and be affected by) bear management. The intensity and location(s) of and

strategies for bear management can have direct effects on land use that may, in turn,

20

affect bear habitat. For example, heavily managed fast-growing tree species (e. g., pine
plantation) may be favored for timber production. Clearcuts can result in an extremely
rigid boundary between highly suitable and unsuitable habitat, and bear managers may
prefer selective harvest of particular size stands that maintain growth of hardwood mast-
producing species (Weaver et al. 1990). Human support for bear management policies
may be affected by personal attitudes toward black bear, which may result from
knowledge (or lack thereof) about bear, or concern about how bear management policies
may affect individual human land use. Land use decisions may affect local economies.
For example, federal mandates for bear habitat conservation may prohibit particular
industries (e.g., timber management, petroleum extraction) from implementing a
preferred management practice. Prevention of an industry from carrying out business
may cause angst among local residents, who depend daily on local industry. As a result,
local residents may be less likely to invest income or other resources into species
conservation on their land.

Completion of this research is critical to the study of wildlife, species
reintroductions, and nature reserves because it (1) develops a systems model to integrate
data of different disciplines that are important to successful wildlife and land use
management, and (2) demonstrates an innovative method for application of spatial
technologies (e.g., remote sensing data) with system modeling tools to integrate, analyze,
and simulate complex relationships that exist between humans and nature. Land cover
and land use data spatially illustrate the relationships of human activity in relation to
potential bear habitat. The survey data provide insight into specific mechanisms of

human activity in relation to bear habitat. Linking these data through use of a systems

21

model increases the ability to understand how humans both directly and indirectly affect
wildlife habitat. Texas Parks and Wildlife Department can use the systems model to
refine management goals, plan public information forums, and ascertain the potential for
bear reintroduction and management over the long term.

Southeast Texas provides a unique opportunity to evaluate a scenario in which it
is known a priori that human activity will be a major influence on a potential black bear
reintroduction and a protected area that may offer refuge to bears. By knowing this from
the outset, human factors can be incorporated into bear management and planning. The
relationships between humans and wildlife habitat that are found in southeast Texas are
not unique to this location. As human activity expands into previously rural locations
worldwide, innovative methods for evaluating changes in both humans and wildlife are

necessary to meet the needs of wildlife while also considering changes in human activity.

22

 

10km

Figure 1.1. Spatial distribution of Big Thicket National Preserve’s administrative units.

23

Socioeconomics (including

attitudes and opinions)
I

Land use - Humans - Demography

 

 

 

 

Intensity
L

Strategy —J Management
U

Locations ‘\\‘

 

 

 

 

 

Quaptity

 

Bear habitat r—Quality

I
Distribution

 

 

 

Figure 1.2. Conceptual framework for determining the feasibility of a Louisiana black
bear reintroduction in and around Big Thicket National Preserve, Texas. Solid arrows
indicate relationships considered in this research. The dashed arrow indicates a
relationship that will likely be more pronounced if and when bears are reintroduced.

24

CHAPTER 2

SCALE, CONTEXT, AND HABITAT ASSESSMENT FOR ECOLOGICAL

FEASIBILITY OF A SPECIES REINTRODUCTION

In collaboration with

Brian A. Maurer and J ianguo Liu

25

Abstract

Accurate estimates of habitat area are necessary for planning wildlife
reintroductions to ensure that adequate resources are available for reintroduced
individuals and their offspring. However, both the scale of analysis used and the context
of habitat within the surrounding landscape can greatly affect estimates of habitat area
and distribution. We incorporated variation in scale of analysis, as well as the
distribution of habitat within the surrounding landscape, to assess the feasibility of a
Louisiana black bear (Ursus americanus luteolus) reintroduction in and around Big
Thicket National Preserve (BTNP), Texas. Analysis of land cover imagery revealed an
estimated 1.3 million hectares of highly suitable habitat across the landscape, and 90% of
land within BTNP was considered highly suitable. However, 44 times more highly
suitable habitat existed beyond the preserve’s boundaries. Average patch size was larger
at a l-ha scale (more appropriate for assessment of bear ecology) than at a 30-m scale
(scale of raw image data) as a result of consolidated small habitat patches at the more
coarse scale. Spearman rank correlations indicated a negative correlation between patch
area and nearest neighbor distance between highly suitable habitat patches. Core habitat
(highly suitable habitat >1 km from human activity) was only visible at the l-ha scale,
and existed primarily within southeastern and northeastern portions of the study area, as
well as within BTNP. Considering both highly suitable and core habitat at the l-ha scale.
we estimated a potential population size of 290 bears within BTNP, and 2,256 for the
whole study area. We believe that a bear reintroduction is ecologically feasible,

particularly if land beyond the preserve is considered.

26

2.1. Introduction

While the area of habitat for many large carnivore species is decreasing, there is
increasing recognition that these animals are valuable members of the ecological
community (Sillero-Zubiri, C., & M.K. Laurenson 2001). As a result, efforts are
increasing to recover populations of locally extirpated species (Reading and Clark 1996).
As human development rapidly expands into previously rural areas, reserves and other
public lands are playing a more important role than public land alone as habitat for
species conservation and recovery initiatives (Woodroffe 2001a). Unfortunately, because
reserves are often not large enough to adequately protect the species for which they were
established (Liu et al. 2001; Gurd et al. 2001; Gurd and Nudds 1999; Soulé and
Simberloff 1986), recovery is often difficult for larger species requiring large and
contiguous areas of habitat, which may no longer exist as a result of human activity or are
currently under private ownership (Pelton 1986; Meffe and Carroll 1996; Woodroffe
2001b)

A particular concern for reserve managers is conflict with humans when wildlife
travels beyond reserve boundaries (i.e., beyond “islands of habitat;” Newmark 1985;
Newmark 1995; Woodroffe and Ginsberg 1998; Hansen and Rotella 2002). Knowledge
about interactions between wildlife and the landscape surrounding a reserve are essential
for management of the reserve and its species (Harcourt et al. 2001). The idea of cross-
boundary management for reserves is well documented in the conservation planning
literature (Pickett and Thompson 1978; Schonewald-Cox and Bayless 1986; Means and
Greene 1987; Schonewald-Cox 1988; Schonewald-Cox et al. 1992; Shafer 1999; Hamin

2001; Harcourt et al. 2001). However, few studies have quantified in detail the

27

magnitude of difference that cross-boundary management can potentially make in
assessing habitat for a particular species. Even more challenging than managing already
existing species is predetermining habitat for use by a reintroduced species. For a
reintroduced species, if habitat is overestimated resources may be too few to meet
population needs. On the other hand, it is possible that underestimating resources may
lead to unexpected overpopulation of a species (Hamilton 1999).

To determine ecological feasibility of a reintroduction, knowledge of two major
items is essential for estimation of a potential target population: (1) the quantity and
quality of available habitat that exists, and (2) distribution of the habitat within the
landscape (Reading and Clark 1996). Because many ecological factors are not easy to
measure, even estimates of habitat area may vary dramatically depending on analysis
tools used and selection of parameters for estimation (Miller et al. 2004; Osborne et al.
2001). In this study, we sought to determine whether methods of analysis (in regard to
scale and context) would lead to varying estimates of habitat quantity and distribution for
a proposed species reintroduction.

Scale is a spatial or temporal dimension characterized by grain (smallest
quantitative unit of area analysis) and extent (e. g., size of the study area; Turner et al.
2001). In recent years, many studies have challenged the issue of how to determine the
best scale of analysis for a species (Norris et al. 2002; Mazerolle and Villard 1999;
Peterson and Parker 1998). For example, Bailey et al. (2002) suggested that distributions
of birds and mammals appeared correlated at a scale related to the landscape level,
whereas plants and butterflies were more sensitive to variables at scales associated with

patches. In fact, many studies have concluded that ecological analysis must take place

28

across many scales (Karl et al. 2000; Stern 1998). Researchers have also suggested
appropriate scales for studying ecological characteristics of species, such as those cited
above, or have explored the effects of changing scale of analysis (Turner et al. 1989).
Because many researchers from different disciplines (e. g., wildlife management, forestry)
may be involved with a species reintroduction, each may unintentionally use different
scales of analysis for different types of data (e. g., 30m scale for land cover data versus
per-hectare number of a particular resource important for a species). Knowledge about
how differences in scale of analysis may greatly affect measurement of habitat for a
species targeted for reintroduction.

Less emphasized in the ecological literature than scale, but also important, is
context. Context is the consideration of surrounding landscape characteristics in relation
to, and how they affect, the feature of interest. Past research has identified the
importance of context, for example, in estimating bird densities among forest fragments.
wetlands, salt marshes, and as related to nest predation (Shriver et al. 2004; Brotons et al.
2003; Riffell et al. 2003; Donovan et al. 1997). Effects of landscape characteristics on
habitat are also of interest in relation to locations of varying land juristicitions. For
protected areas, landscape context may involve isolated areas of habitat that are
surrounded by an inhospitable human-modified matrix (Gilpin and Diamond 1980).
which is logically equivalent to the issue of cross-boundary management.

Through use of land cover and habitat analysis, we applied the ideas of habitat
area estimation and cross-boundary management to study how scale and context affect
estimates of available habitat for a reintroduction. Our goal was to estimate the

ecological feasibility of a Louisiana black bear (Ursus americanus luteolus)

29

reintroduction in and around Big Thicket National Preserve (BTNP), Texas. In the
southeastern United States, black bear habitat includes hardwood, mixed hardwood-pine
forest, and bottomland areas that contain a variety of mast-producing tree species (e. g.,
oak, hickory; Dobey et al. 2005; Larkin et al. 2004; Maehr et al. 2003; BBCC 1997;
Bowker and Jacobson 1995). Older-growth forests in remote areas (i.e., core habitat, or
locations >1 km from human activity; Linnell et al. 2000) are more likely than younger
forest to contain large trees for denning (Lariviere 2001; Pelton 1982), although southern
bears often den in slash piles, excavated ground tree cavities and ground depressions
(Hellgren and Vaughan 1989b; Weaver et al. 1990). Early successional undergrth
containing many fruit-bearing herbaceous species is considered high quality habitat
(Anderson 1997; Marchinton 1995; Nyland 1995; Hellgren and Vaughan 1991). Specific
objectives were to: (1) quantify how much habitat is available within the preserve, and
how this estimates of habitat changes when the preserve is considered within the context
of the surrounding landscape, (2) estimate changes in habitat area as a result of changing
the scale of analysis, (3) assess how changing the scale of analysis affects estimated
connectivity of habitat, (4) identify and quantify core areas (see Methods) of habitat, and
whether their distribution change as scale of analysis changes, and (5) based on habitat

area, estimate black bear population size that may exist within the study area.

2.2. Methods
2.2.1. Background
The Louisiana black bear once existed throughout all of Louisiana, southeast

Texas, and southern Mississippi (BBCC 1997). With one of the historically densest

30

populations of bears, early settlers of the region relied on bear as a source of food (Truett
and Lay 1984). Overharvest and habitat destruction led to near demise of this subspecies
during the early 19005 (BBCC 1997). During the second half of the 20th century, only
two small remnant populations of the subspecies existed, both in eastern Louisiana. An
extensive public outreach campaign during the 19905 led to recovery efforts for
Louisiana black bear in Louisiana, as well as the creation of two management plans that
focused on recovery throughout the species historical range (BBCC 1997; Bowker and
Jacobsen 1995). Thus far, recovery in Louisiana has been successful (Van Why, 2003)
and feasibility analyses have been completed for Mississippi (Bowman 1999; Shropshire
1996).

In eastern Texas, the number of black bear sightings has increased during the past
decade (N. Garner, Texas Parks and Wildlife Department (TPWD), personal
communication). Because no breeding population of bears is known to exist in Texas, it
is likely that these bears are transients from Arkansas, Oklahoma, or Louisiana that
wander into Texas. The sightings have prompted creation of an East Texas black bear
management plan. Objectives of the management plan for the next 15 years include
public coordination, communication, outreach and information dissemination, habitat
management, and research (TPWD 2005). The ultimate goal is to restore habitat for the
purpose of reestablishing black bear as a viable ecosystem component in southeast Texas

(TPWD 2005).

31

2. 2. 2. Study area

Our study area consists of the 39,285-ha BTNP in southeastern Texas (Lat/Long.
—94.36/30.38; Figure 2.1) and 12 counties that include and surround the preserve. BTNP
is unique in that it consists of 12 disjunct management units (Figure 1.1), seven of which
are river corridors that connect larger non-corridor units. This distribution makes BTNP
an ideal location in which to study human impacts on wildlife and nature reserves.
Approximately 500,000 residents live within the study area (US Census Bureau data).
Much of the area is rural, with numerous small towns throughout and one larger
community (City of Luflcin) along the northwestern boundary. More than 75% of the
land is managed for timber or owned by the Federal government (i.e., BTNP and four
national forests). The southern edge of the study area, which is also the most densely
populated location, consists of suburban development from the cities of Houston and

Beaumont.

2. 2. 3. Habitat classification

Land cover classification was derived from 2002 LandsatTM ETM+ (30m
resolution) multispectral imagery. We selected images from November and March to
maximize availability of cloud-free images, and selected training sites (i.e., ground-
reference points from which land cover data were collected) that would be easily and
consistently identifiable on imagery through all seasons. During the summers of 2002
and 2003, used a GarminTM hand-held global positioning system (GPS) to ground-truth
255 training sites to use as a basis for supervised imagery classification. Applying these

data to imagery classification, we identified seven general land cover categories: urban

32

(including residential), water, sand, agriculture (including livestock pastures), pine forest
(both plantation and natural), mixed (pine and hardwood) forest, and bottomland forest
(hardwood forest within bottomland areas; adapted from Harcombe and Callaway 1997).

To identify black bear habitat, we recoded land cover data to derive four habitat
classes: highly suitable, suitable, marginal, and unsuitable. Assignment of each of the
seven land cover classes to one of the four specific habitat classes was based on past
research on black bear ecology for the south-central US (Larkin et al. 2004; Maehr et al.
2003; Marchinton 1995; Weaver et al. 1990; Wagner 1990; Pelton 1982). Highly
suitable habitat included mixed and bottomland forest. Common species within mixed
forest included oaks (Quercus spp.), hickory (Carya spp.), magnolia (Illicium spp.),
beech (Fagus spp.), and pine (Pinus spp.). It is believed that beech-magnolia—loblolly
pine (P. taeda) was the historical climax community of the region (Peacock 1984; Marks
and Harcombe 1981; Harcombe and Marks 1977) before many areas were replaced with
faster regenerating pine plantations. Bottomland forest included cypress (Taxodium
spp.), American holly (Ilex opaca), tupelo (Nyssa spp.), sweetgum (Liquidambar
styraciflua) and lowland oak, hickory, maple (Acer spp.), birch (Betula spp.), sycamore
(Platanus occidentalis), and ash (Fraxinus spp.) species. Mast and fi'uits related to these
forest types may be important food sources for bears.

Suitable habitat included recent clearcuts containing downed slash and young
regenerating forest. These areas contained thick layers of early successional vegetation
and saplings dominated by honeysuckle (F. Caprifoliaceae) and poison ivy
(Toxicodendron radicans). Throughout the south-central US, bears use slash within

regenerative areas for denning, and corresponding vegetation would be among prime

33

foraging habitat (Larkin et al. 2004; Anderson 1997; Marchinton 1995; Nyland 1995;
Wagner 1990; Weaver et al. 1990). If managed for seedling species often present in
natural regeneration (e.g., oak), these areas may become highly suitable in the future until
the next scheduled harvest. Although regenerating forest and clearcuts may contain
favorable vegetation and denning sites for bears, we classified such areas separately from
older growth forest because of uncertainty related to management.

Sandy areas and pine forest were defined as marginal habitat. Sand included
sandbars along rivers and areas of arid-environment plants (e. g., yucca, cactus; Peacock
1984). Bears may forage among these plants, but are not likely to spend large amounts of
time in such sun-exposed areas (Weaver 1992). Pine included slash pine (Pinus elliottii)
plantations, as well as longleaf pine (Pinus palustris; Peacock 1984), that were managed
for removal of deciduous undergrowth.

Agricultural and urban land cover were designated as unsuitable habitat. Only
along the southern boundary of the study area (i.e., Gulf Coastal Plain) was row crop
agriculture prevalent, and consisted of millet and rice among other products. Throughout
the remainder of the study area, row crop agriculture was sparse and livestock rearing
was widespread. Most livestock pastures were fenced in with barbed wire, which may
ultimately deter bears (J onker et al. 1998). Finally, because specific urban characteristics
(e.g., low-density residential versus high-density residential) were difficult to disseminate
using 30-m resolution imagery, we classified all identifiable human infrastructure (e. g.,

houses. buildings) as urban. We removed large water bodies of water from analysis.

34

2. 2. 4. Habitat analysis

Using Imagine 8.7 (Leica Geosystems GIS & Mapping, LLC, St. Gallen,
Switzerland), we grouped contiguous pixels of the same land cover class to represent
patches of a similar habitat type. To account for classification errors and ecological
characteristics of black bears, we adapted imagery analysis to simulate landscape analysis
at different scales. First, unfiltered data were used to simulate a 30-m (0.09-ha) scale,
equivalent to the grain of unmanipulated image analysis.

Second, we assessed habitat at a scale relevant to bear ecology and behavior. We
applied a moving window filter to image data to reclassify each image pixel based on the
most common habitat class of surrounding pixels. This method was similar to placing a
grid of a particular grain over the landscape and applying the most common land cover
within a particular quadrant to the quadrant (Chapin et al. 1998). The results simulated
patches of particular habitat classes at a predetermined scale. We determined scale of
analysis based on information from previous studies of black bears to incorporate land
cover data into habitat estimates. Estimation of habitat at the scale of a black bear home
range would be ideal, but individual bear home ranges can range from fewer than five to
hundreds of square kilometers (Garshelis and Pelton 1981; Pelton 1982; Hellgren and
Vaughan 1990; Marchinton 1995; Rudis and Tansey 1995; Wooding et al. 1994b;
Wagner 1990; Anderson 1997; Beausoleil 1999; BBCC 1997; Lariviére 2001; Maehr et
al. 2003), and vary by season (Garshelis and Pelton 1981; Hellgren and Vaughan 1990).
Within fragmented forest areas of eastern Louisiana, bear home ranges within the
Atchafalaya River Basin were approximately 340 km2 for males and 42 km2 for females

(Marchinton 1995). and 33 km2 and 15 km2 for males and females, respectively with the

35

Tensas National Wildlife Refuge (Wagner 1990). In exploratory analysis, we
manipulated imagery to simulate habitat at the scale of one kilometer, and observed the
elimination of several features (e.g., entire small towns) that are likely important for
determining habitat distribution. Therefore, we resolved that our selection of scale
should be based on distances of bear movement rather than home range.

A majority of bear movement estimates are based on estimated average
interlocation distance from radiotelemetry observation, which typically occurs at intervals
of one hour to 24 hours depending on the method of data collection (i.e., aerial versus
ground; Garshelis and Pelton 1981; Hellgren et al. 1991; Marchinton 1995; Nyland 1995;
Wagner 1990). Average movement-related interlocation distances from past southeastern
black bear studies have ranged from several hundred meters to more than five kilometers,
with females often moving shorter distances than males (Reynolds and Beecham 1980;
Marchinton 1995; Wagner 1990; Eastridge 2000). Basing our selection on a conservative
daily movement distance (0.25 kin/24 hour period; Marchinton 1995; Wagner 1990;
Eastridge 2000), we adapted a comparative scale of analysis of 0.01 km, or
approximately movement of within one hectare per hour. It is important to note that we
divided movement distance across a 24-hour period because bear activity varies
depending on time of day, and several hours are spent resting (Garshelis and Pelton 1981;
Pelton 1982; Lariviere 2001). For both scales of analysis (i.e., 30-m/0.09-ha and l-ha).
and all four habitat classes (i.e., unsuitable, marginal, suitable, and highly suitable), we
made three habitat calculations: (1) total area of each habitat class throughout the study
area, (2) area of habitat within the preserve, and (3) total area of habitat outside of the

preserve.

36

Rudis and Tansey (1995) suggested at least 7,580 ha are necessary for a
population of 50 bears (or approximately 150 ha/bear). However, Marchington (1995)
suggested denser bear populations within fragmented bottomlands of Louisiana of
approximately 50 ha/bear. Similar to bear habitat in Louisiana, BTNP and the
surrounding area contain fragmented land cover with heterogeneous land use. We
evaluated all highly suitable habitat patches >50 ha across the study area, and identified

“patch” as a contiguous area of the same habitat classification (Turner et al. 2001).

2. 2. 5. Habitat connectivity

Several researchers have addressed the importance of habitat connectivity for
species movement and dispersal (Fahrig and Merriam 1985; With et al. 1997; Wiegand et
al. 1999), and appropriate measures of connectivity (Schumaker 1996; Tischendorf and
Fahrig 2000a,b; Moilanen and Hanski 2001; Turner et al. 2001). Black bears have been
known to travel large distances within their home range both across areas of perceivably
unsuitable habitat, as well as through woodland corridors (Marchinton 1995; Wagner
1990). Therefore, although we are uncertain as to whether connectivity of highly suitable
habitat will be important for bear movement in and around BTNP, we quantified
connectivity of highly suitable habitat.

We used Patch Analysis 3.1 (Elkie et al. 1999) to determine connectedness of
highly suitable black bear habitat both within BTNP and across the study area. Because
of the large size of land cover imagery files and resulting processing difficulties, we
divided the study area into equal sized hexagonal polygon regions. After experimenting

with polygon regions of various sizes, we selected polygons (440 total) with an area of

37

1000 ha that allowed for both consideration of large habitat fragments that may be
important to bear (Rudis and Tansey 1995; Linnell et al. 2000) and image processing
within a reasonable amount of time. We hypothesized a negative relationship between
highly suitable habitat per hexagon and habitat connectivity. In other words, as area of
highly suitable habitat per hexagon increased, the more likely that habitat patches would
be closer together with a smaller mean nearest neighbor distance. Assuming that suitable
habitat may also be important for bear refuge and foraging, we completed connectivity
analyses for highly suitable and suitable habitat, as well as a combination of both.

Using ArcView GIS 3.2 (Environmental systems Research Institute, Inc.), we
selected mean nearest neighbor distance to measure connectivity of bear habitat. Mean
nearest neighbor distance is a measure of patch isolation, and calculated by the average
nearest neighbor distance for habitat patches within each hexagon (Elkie et a1. 1999).
After plotting the data, which revealed that it did not meet assumptions of normality, we
applied Spearman rank-order correlations to assess relationships between habitat area and
nearest neighbor distance (Dytham 2003). Significant results at both the 90% and 95%

were reported.

2. 2. 6. Core area analysis

Results from past studies of black bear ecology and management have suggested
that presence of large areas of remote forest (core habitat) is an important feature of black
bear habitat (BBCC 1997; Pelton 1982). Rudis (1986) defined “remote” as forest at least
0.8 km from human activity. Core habitat has also been described as a contiguous forest

tract >1 ,000 ha (Rudis and Tansey 1995). Linnel et al. (2000) suggested that bears

38

tolerate human activity that is at least one kilometer away from dens. Therefore, we
identified remote as highly suitable habitat at least one kilometer away from human

activity (unsuitable habitat), based on the Euclidian distance from unsuitable habitat.

2. 2. 7. Estimation ofpotential black bear population size

In the southeastern United States, estimated black bear population densities have
ranged from 0.008 — 3.6 bears/ka (Smith 1985; Wooding et al. 1994b; Hellgren and
Vaughan 1989a; Marchinton 1995; Rudis and Tansey 1995; Wagner 1990). The densest
estimates were found within the Tensas River Basin, which consists of fragmented forest
surrounded by agriculture fields (Marchinton 1995). Because these estimates vary
greatly, we developed an estimated population size for a range of densities, from a
conservative 0.005 bears/km2 to 3.6 bears/km2 at intervals of 0.05 bears/kmz. Assuming
a 1 :1 sex ratio, estimated population size at each density was a factor of total highly
suitable habitat area divided by hypothetical bear density. Because our core area analysis
focused on areas >1 km away from human activity, we also re-calculated estimated
potential bear population capacity based on area of highly suitable habitat that was >1 km

away from human activity.

2.3. Results
2.3.]. Estimation of total potential habitat area inside and outside the preserve

At the scale of image classification (30 m), the mixed forest was the land cover
class with the greatest area, occupying approximately 40% of the study area (Table 2.1).

Pine (11.68%), clearcut/regeneration (18.42%), and bottomland forest (9.52%) each

39

covered approximately 10% or more of the study area. Urban, grass/pasture, sand, and
water all contained <10% of the area.

Water existed mainly within three large reservoirs (Sam Rayburn Reservoir,
Toledo Bend Reservoir, and Lake Livingston). Eliminating water from further analysis,
area of the seven remaining land cover classes translated to 391,952 ha unsuitable,
310,393 ha marginal, 485,981 ha suitable, and 1,296,101 ha highly suitable habitat. Of
the total land area (water eliminated; 2,484,427 ha), this was equivalent to 15.78%,
12.49%, 19.56%, and 52.17% respectively. Most of the unsuitable habitat was located
along the southern and northwestern portions of the study area (Figure 2.2). Marginal
and suitable habitat classes were well distributed, with concentrations of both in eastern
and central portions of the study area. Highly suitable habitat was scattered, with large
concentrations within the south central, western, and central portions of the study area.

A majority of area within BTNP was identified as highly suitable habitat (Table
2.2), and the remaining three habitat classes each consisted of <10% of preserve area.
Excluding BTNP, approximately half of the study area was identified as highly suitable,
with each of the remaining habitat classes making up <20% of the remaining area. Total
area of all habitat classes as well as percent of unsuitable, marginal, and suitable habitat
were approximately the same across the whole study area as compared to outside of the

preserve (Table 2.2).

2. 3. 2. Estimation of habitat area based on scale of analysis

Area of each land cover class varied based on scale of analysis (Table 2.3). Area

of urban, pine, sand, and bottomland forest classes decreased as scale increased.

40

Grass/pasture and mixed forest area increased as scale decreased. Clearcut/regeneration
and water decreased slightly as scale increased. For habitat classification, area of
unsuitable, marginal, and suitable habitat decreased as scale of analysis increased,
whereas area of highly suitable habitat increased as scale of analysis increased.

At the scale of 1 ha, 90% of land within BTNP was classified as highly suitable,
and increase in percent area from the 30m scale. Area of suitable habitat decreased to six
percent of the total. Marginal and unsuitable habitat decreased to less than three and one

percent, respectively, of total land within the preserve.

2. 3. 3. Variation in patch quantity and size across scales

Excluding all patches with <50 m in area, average patch size was larger at the 1 ha
scale than at the 30m scale (Table 2.4). Median patch size was similar across both scales.
which is likely a result of, although large patches were unchanged, smaller patches were
consolidated into larger patches. Within each scale of analysis, marginal and suitable
habitat contained the on-average smallest patches, whereas highly suitable habitat
contained the on-average largest patches.

Frequency of habitat patches of various areas differed between scales (Figure 2.3).
Variation in the number of patches of each size class existed between habitat types, but
differences were insignificant based on the overlapping of standard error bars. In
general, the total number of patches for a particular interval increased as scale increased
for marginal and suitable habitat, and minimally for unsuitable habitat. However, the

opposite relationship existed for highly suitable habitat. Highly suitable habitat contained

41

a greater frequency of larger patches, whereas suitable habitat contained a greater

frequency of smaller patches.

2. 3. 4. Habitat connectivity

Spearman rank correlations for habitat connectivity varied between habitat classes
and scales (Table 2.5). At both scales, suitable habitat within BTNP, combined
suitable/highly suitable habitat within BTNP, highly suitable habitat across the whole
study area, and combined suitable/highly suitable habitat across the study area had
significantly negative correlations with mean nearest neighbor distance. However, highly
suitable habitat within BTNP at both scales, suitable habitat across the study area at the
30m scale, and combined suitable/highly suitable habitat at the 1 ha scale were positively
correlated with mean nearest neighbor. Therefore, greater areas of habitat were more

likely to be located closer together.

2.3.5. Distribution of core habitat area

Presence of core habitat varied between scales. There was complete lack of core
habitat at the 30 m scale, even within BTNP (Figure 2.4). However, large areas of core
habitat existed at the 1 ha scale, totaling approximately 225,626 ha. These core areas
existed primarily within the southeastern and northeastern portions of the study area,

along the eastern boundary of the study area, and within BTNP.

42

2. 3. 6. Estimation of potential black bear population size

Assuming a direct relationship between population size and habitat area, estimates
of the potential black bear population size across the study area varied based on scale of
analysis. At a density of one bear per square kilometer and considering all available
highly suitable habitat, estimated bear population sizes are approximately 12,000 and
15,000 at the 30 m and 1 ha scales, respectively. However, these estimates are likely too
high. With an estimated 29,000 hectares of highly suitable habitat (290 krnz) within the
preserve, we estimated that BTNP offers enough habitat for approximately 290 bears,
with a density of I bear per square kilometer, but we reiterate that this was a total
estimate for all 12 disjunct preserve management units. When potential bear population
size was recalculated for the 225,626 ha of core area only, at a density of 1 bear/kmz, the
result was an estimated 2,256 bears for the entire study area, and 290 within the preserve.
This is likely a more reasonable population estimate for bear management goals. as it

takes into account both area of highly suitable and core habitat.

2.4. Discussion

Most of the land within the preserve (81%), and half (50%) of the land outside of
the preserve was classified as highly suitable at the scale of unmanipulated imagery (30
m). Based on the size of BTNP (39,256 ha, or 392.6 kmz), the large concentration of
highly suitable habitat within the preserve may provide adequate refuge for
approximately 290 black bears. However, it is important to reiterate that BTNP’s 12
units are disjunct and, therefore, the preserve cannot be thought of as one contiguous unit

of habitat. Black bear home ranges and habitat use can vary widely based on sex,

43

location, season, and distribution of resources (Lariviere 2001; Hellgren and Vaughan
1991; Pelton 1982; Garshelis and Pelton 1981). Although estimates of black bear
densities vary widely (see BBCC 1997 for a summary), assuming a 1:1 sex ratio and a
density of 1 bear/kmz, we estimated a potential population size of approximately 290
bears in the preserve. However, this was likely an overestimation because the disj unct
units of BTNP range from 58 ha (likely too small for even one black bear) to 11,000 ha.
In other words, even though a majority of the preserve contains highly suitable habitat,
each individual preserve unit alone may not be large enough for viable black bear
population. As a result, we expect that bears will travel beyond the preserve boundaries
for foraging and mating. This may not be considered ideal for nature reserve design
theory (Higgs 1981; Diamond 1975), but the reality of this situation emphasizes the
importance of considering landscape context.

Considering the preserve within the context of the surrounding landscape, there is
an estimated 1.3 million ha of highly suitable habitat. This is greater than 44 times more
habitat than within the preserve alone, but less forested area than estimated by Rudis and
Tansey (1995) ten years prior. Estimated potential black bear population size for the
whole area then becomes much greater than 290 bears (Rudis and Tansey 1995; Hellgren
and Vaughan 1989a) for two main reasons besides the simple fact that we are considering
a much larger area. First, black bear home ranges can overlap by 20-50% throughout the
year; females, particularly related individuals, may overlap exclusively (Hellgren and
Vaughan 1990; Marchinton 1995). Second, we considered the distribution of only highly
suitable habitat for population estimation. If bears use suitable or marginal habitat, then

estimates of habitat and potential population size should be reevaluated. This may be

44

necessary, as colonizing individuals of species, including dispersing bears (Rogers 1987),
often explore a larger area of habitat than individuals within a larger population
(Diamond 1975). Regardless, because no data are available for comparing bear
movements with estimated habitat for our study directly, use of our estimates should be
made with caution and be re-evaluated as movement data become available.

Many areas of highly suitable habitat run along rivers, and may provide
connectivity between preserve units and other large patches of highly suitable habitat. In
fact, four of the BTNP units (Figure 2.1) are preserve-managed river corridors that link
larger preserve units. Positive relationships between area of highly suitable habitat and
nearest neighbor distances for BTNP land were likely an artifact of the preserve’s
distribution. Results from our connectivity analyses suggested that across the study area,
habitat connectivity increased as the amount of highly suitable habitat per unit area
increased. Research that has explored the utility of habitat connectedness by wildlife
corridors has been met with mixed results (Haddad 1999; Beier 1993; Simberloff and
Cox 1987). Results from Beausoleil (1999) and Anderson (1997) suggested that bears
within a fragmented landscape in Louisiana showed preference for forest fragments and
corridors and avoidance of agricultural fields. In fact, >50% of bears from the same
studies used corridors for movement between forested tracts. However, Marchinton
(1995) suggested that bears often ventured across open agricultural fields rather than
remaining within the cover of forested corridors. Unfortunately, bear telemetry data were
not available for our study, but we suggest that wildlife managers maintain habitat along
river channels and among private lands between units of BTNP for monitoring bear

movement. It is also important to note that because of the coarse analysis of land cover

45

used, finer landscape features (e.g., roads) were not considered beyond identification
through imagery classification and must be further evaluated. Therefore, to address our
first and fourth research questions, when considering the preserve within the context of
the larger landscape, there is adequate habitat for a black bear population and we
conservatively estimated a potential population of >8,000 bears based on an average
density of one bear/kmz. However, initial population estimates should be made with
caution because pioneer individuals may initially use habitat differently than members of
a larger population in the future.

Across the landscape, changing the scale of analysis from 0.09 ha (30-m pixel) to
a l-ha scale resulted in different estimates of available highly suitable habitat across the
study area and within the preserve. In fact, across the study area there was a >12%
(1,730 kmz) increase in highly suitable habitat when moving from the 0.09-ha to the l-ha
scale of analysis, which affected our estimates of potential black bear population size
(Figure 2.5). Therefore, scale of analysis that is adopted by management will affect
estimated habitat (Stern 1998). Regardless of scale, managers must determine through
monitoring whether locations used by bears are consistent across scales. For example,
Chapin et al. (1998) observed that habitat used by adult martin corresponded to grid cells
containing a greater percentage of forested cover and larger forest patches across four
spatial scales. As in our study, however, the overall geographic distribution of habitat
across the study area in Chapin (1998) did not change across scales. An increase in
habitat corresponding to an increase in scale does not hold true for all of our habitat
classifications. For suitable and marginal habitat, the area of habitat increases as scale

decreases from one hectare to 0.09 ha by 12% and 42%, respectively. Area of unsuitable

46

habitat remained relatively the same across scales, which supports the suggestion that
utility of remote sensing imagery for consistent identification is better for some landscape
features than others (Allnutt et al. 2002; Cardillo et al. 1999). As is the case for using
any remotely sensed data, researchers and managers must consider reliability of
classification, and how reliability varies as image data are manipulated to represent
various scales. Aggregating pixels to a more coarse scale than raw (30-m) classification
reduces spatial heterogeneity and how a potential bear population is estimated.
Maintaining a finer scale accounts for landscape heterogeneity and results in a more
conservative bear population estimate. Therefore, estimates of highly suitable habitat
area does change based on scale of analysis. For instance, if managers assess habitat at a
scale of l-ha, but habitat analysis by researchers uses a 0.09-ha scale, then total area of
highly suitable habitat will be underestimated, whereas suitable and marginal habitat will
be overestimated. Regardless, when selecting a scale for analysis, researchers and
managers must seriously consider drawbacks of manipulating image data (e. g., pixel
aggregation) when selecting a consistent scale for analysis.

Average and median patch size increased as scale of analysis increased for all
four habitat classes. This is likely because more coarse scales resulted in the combination
of similarly classified pixels into larger groups. Because the resolution of this imagery is
30-m, the lower possible range of patch sizes will, using a smaller scale, result in a larger
number of smaller patches than what will be determined at a larger scale. Many single
pixels of a particular habitat classification become reorganized into a more prominent

habitat type.

47

As distance from the nearest unsuitable habitat increases, the total area of highly
suitable habitat increased drastically, and then decreased. For both scales, the greatest
area of habitat existed at a distance of 150 m from the nearest unsuitable habitat. The
distribution is relatively the same for the l-ha scale, both quantitatively and spatially.
Existing remote areas correspond with location of BTNP, National Forests, and river
corridors. At the 30-m scale, public lands are identified as further from unsuitable habitat
than other locations across the study area, but no remote areas exist. Similar to change in
patch size across scales, many erroneous single pixels of unsuitable habitat are
incorporated into larger patches of other habitat types when the data are clumped to
simulate the larger scales. Therefore, the identification and quantification of core areas
and whether their existence is consistent across scales, existence of core areas (defined
according to results from Linnell et al. 2000) depends on scale of analysis. For the l-ha
scale, core areas exist mainly in relation to distribution of public lands and along river
corridors. However, focusing on core areas only could only be completed at the l-ha
scale, and the estimated potential population size of bears decreased greatly to 2,256
bears. This is a more accurate measure of the region’s bear carrying capacity as there are
likely factors that will affect bear populations that we did not consider in this study (e.g.,
annual mast crop output).

The implications of these results as related to ecological feasibility of a black bear
reintroduction in southeast Texas are several. First, the disj unct distribution of BTNP
may be considered spatially well done by general reserve planning recommendations
(Diamond 1975). Enough highly suitable habitat existed within BTNP for a small black

bear population (potentially as many as 290 bears). However, because of the preserve’s

48

small size, disjunct distribution, and as-discussed habitat needs of black bears, BTNP
cannot be thought of as a single large preserve that can self—sustain a black bear
population. With 44 times more habitat existing beyond the preserve’s boundaries, bear
managers must assume that activities beyond the preserve’s boundary will be used by
bears and critical to a successful bear recovery (Parks and Harcourt 2002). Because the
goal of this study was to estimate the area of potential habitat, whether island
biogeography theory (Diamond 1975) is sufficient for meeting bear habitat-area needs
will not be determined if and when bears are released. If the preserve is considered
within the context of the surrounding landscape, adequate area of highly suitable habitat
exists for a viable bear population.

Second, highly suitable habitat is well distributed across the study area. River
channels, particularly those that are within boundaries of the preserve, may become
critical habitat that bears use for foraging, denning, and moving between preserve units.
Where highway overpasses exist over rivers, river channels may become important as
underpasses under highways for bears moving from location to location. Such was the
case in Florida, where bears often used highway underpasses within their home range
(Foster 1995). However, rivers themselves may be impediments to bear movement
(White et al. 2000). As mentioned, additional fine scale analysis is necessary to
determine where roads and other individual elements of human activity may hinder bear
movements.

Third, large patches of highly suitable habitat exist, but the area of highly suitable
habitat decreases with increasing distance from unsuitable habitat. Linnell et al. (2000)

suggested that bears selected dens >1 km away from human activity and tended to not

49

tolerate activity at less than this distance. Use of the l-ha scale allowed us to identify
areas of highly suitable habitat that meet this criterion. Such remote areas were
distributed mainly among public lands and river channels, and it is important to note that
presence of remote areas disappeared entirely at the 30—m scale of analysis. Therefore,
adoption of a l-ha scale of analysis and a more conservative estimate of a potential bear
population (2,256 bears) may be more appropriate for initial bear recovery goals. When
conducting fine scale analysis, managers should make note of unsuitable habitat at the
0.09-ha scale that is eliminated at more coarse scales of analysis, and determine whether
these locations offer infrastructure that may threaten the well being of bears that move
beyond core habitat areas.

From our coarse scale analysis, and when the preserve is considered within the
context of the surrounding landscape, we believe that a black bear reintroduction is
ecologically feasible, but additional fine scale analysis may uncover elements within the
study area that must be considered in management planning. However, the landscape of
southeastern Texas will likely experience major changes that present threatening
uncertainty to the quantity and quality of black bear habitat. Much of the private land
throughout the study area is owned and managed by timber companies (see Chapter 2).
As a result of a land divestment of >600,000 hectares by one timber company, large areas
of habitat may be lost to human development. Proposals have been made to build water
management structures to supply water to Texas’s growing cities. In addition, plans have
been suggested to build an eight-lane superhighway that would bisect units of BTNP, and
could potentially act as a barrier to wildlife movement between preserve units.

Knowledge about the distribution of bear habitat may aid resource managers and

50

development coordinators to meet the needs of both regional wildlife and human

development.

51

Table 2.1. Area and percent of total area of the eight land cover categories derived from
classification of remotely sensed imagery (30-m scale) for estimation of black bear
habitat in and around Big Thicket National Preserve, Texas.

 

 

Land cover classification Total area (1,000 ha) Percent of total area8
Urban 188.3 7.1
Grass/pasture 203.7 7.7
Pine 308.2 1 1.7
Sand 2.2 <0.1
Clearcut/regeneration 486.0 18.4
Mixed forest 1,045.0 39.6
Bottomland forest 251.1 9.5
Water 154.3 5.9

 

aTotals >100% are a result of rounding.

52

.wficcze mo :38 a o8 o\ooo~ 9 :33 8c 23o?

 

 

34am 82¢ :8; as: 4.3:. 2.3 33 :d: was D>.~....aa.:oz
EM 6. as com is 3 :8 am :5 S 38 Essa
4.32% 3.8 283 6.2V mam? cam: 32 m as: 8. § 2282.03 as 3% $5
3.5 98— _So._. @3815 3392 038:5 BEwSE 038533 Eme

 

C38 :39 mo H583 a: coo. 3 55¢: («o m8<

 

.AE cmv cosmocfimflo Downs: mo 2QO of 8 mam mo 033:0 was .EZFmV otomocm Ecoumz
“83:: mi :33; .85 33% 20:3 2: $88 2352 83 299:6 bawE can 639:5 .EEwSE JESSE: mo 354 .m.m 2an

53

Table 2.3. Area (and percent change) of land cover classes and corresponding black bear
habitat at 30-m (0.09-ha) and 1 ha scales.

 

 

 

 

Area (1 ,000 ha)
Classification 0.09 ha 1 ha Percent change
Urban 188.3 135.5 —28.0
Grass/pasture 203.7 31 1.4 65.4
Pine 308.8 236.4 —-52.9
Land cover Sand 2.2 1.1 —50.0
Clearcut/regeneration 486.0 471.4 — 3.0
Mixed forest 1,045.0 1,244.2 19.1
Bottomland forest 251.1 186.3 —25.8
Water 154.3 151.4 -— 1.9
Unsuitable 488.9 479.3 — 2.0
Habitat Marginal 310.4 21 1.2 —32.0
Suitable 487.4 430.6 —1 1.6
Highly suitable 1,296.7 1,470.0 13.4

 

54

Table 2.4. Mean (iSD) patch size of black bear habitat at 30-m (0.09 ha) and 1 ha scales.

 

 

 

Scale Habitat category N Mean (iSD; ha) Median patch (ha)
Unsuitable 666 438.2 (3356.8) 97.8
Marginal 291 100.9 (102.2) 73.4

0.09 ha Suitable 1,265 121.9(158.2) 81.3
Highly suitable 1,460 1,593.3 (16,903.7) 116.2
Total 3,682 554.3 (8,547.0) 89.5
Unsuitable 744 506.5 (4,004.1) 97.3
Marginal 448 116.4 (137.7) 78.9

1 ha Suitable 1,627 142.0 (199.9) 87.8
Highly suitable 470 2,984.2 (49,281.2) 114.1
Total 3,289 627.1 (18,7354) 90.6

 

55

Table 2.5. Spearman rank correlation values (r,)3 for nearest neighbor analysis of
suitable and highly suitable black bear habitat connectivity at 30m and 1 ha scales within
Big Thicket National Preserve (BTNP) and across the whole study area.

 

 

 

rs
Habitat class 30 m 1 ha
BTNP — suitable -0.554* -0.374**
BTNP — highly suitable 0.325 0.238
BTNP — suitable and highly suitable -0.380* -0.353*
Whole area — suitable 1.000* -0.567*
Whole area — highly suitable —0.272* -0.460*
Whole area — suitable and highly suitable -0.426* 0.3 89*

 

a"‘Significant at the 0.05 level (2-tailed)
“Significant at the 0.01 level (2-tailed)

56

Figure 2.1. The 12-county region in southeastern Texas used to evaluate land cover and
potential habitat for a black bear reintroduction. These counties include and surround Big
Thicket National Preserve (shaded areas in oval).

57

Figure 2.1.

 

 

San
Augustine _

 

  
 

 

 

Angelina

 

 

 

 

 

 

 

 

    

San
Jacinto

w-z...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

58

 

Figure 2.2. Distribution (shown in black) of (a) unsuitable, (b) marginal, (c) suitable. and
((1) highly suitable black bear habitat throughout the study area in southeastern Texas.
Large water bodies are shaded in light gray.

59

   

 

 

 

 

 

3.50
3.00 1'
:9? 2.50
3 2.00 , , - 77777
8‘
d: 150 , ******* 7 7
OD ‘ .
O
.—3
1.00
0.50
0.00 , 474 m- ‘ " 7 ——
50-99.99 loo—499.99 500-999.99 1000- 5000- >10000
4999.99 9999.99
Area of patch (ha)
(b)
3 , ,
5
:1
3 l
a: i i
an , , ,
O
h] i i
1
50-99.99 loo-499.99 SOC-999.99 1000- 5000- >10000
4999.99 9999.99
Area of patch (ha)

Figure 2.3. Distribution (log frequency) of (a) unsuitable, (b) marginal, (0) suitable, and
(d) highly suitable patches of black bear habitat by area (ha). Confidence intervals were
defined at the 95% level.

60

Figure 2.3. (Cont’d.)

(C)

Log frequency

((0

Log frequency

 

11m

 

50-9999 IOU-499.99 500-999.99 1000- 5000- >10000
4999.99 9999.99
Area of patch (ha)

3.50

3.00

 

,
l .
.1 il

50-99.99 100-499.99 SOC-999.99 1000- 5000- 1‘" 10000
4999.99 9999.99

 

Area of patch (ha)

61

 

62

Figure 2.4. Distribution of core habitat area at (a) 1 ha and (b) 30 m scales, designated by dark gray shaded areas. Note that no core

habitat exists at the 30 m scale, even within BTNP (outline).

CHAPTER 3

IMPORTANCE OF PRIVATE LAND CONSERVATION FOR A BLACK BEAR

RECOVERY

In collaboration with

Jianguo Liu

63

Abstract

Although most land within the US is privately owned, wildlife conservation
efforts have largely focused on public land. In the south-central US, privately owned
timberlands have helped maintain extensive areas forest that are important as wildlife
habitat. The objective of this study was to assess the importance of regional private
timberlands on a Louisiana black bear reintroduction in and around Big Thicket National
Preserve (BTNP), Texas. Since establishment of BTNP in 1974, the overall area of
highly suitable habitat has decreased, as well as the total area of timberlands. The
percentage of highly suitable habitat among timberlands has increased, and, in 2002,
>3 0% of highly suitable habitat across the study area was contained among timberlands.
Furthermore, timberlands contained approximately half of all land within the study area
that has been consistently highly suitable since establishment of BTNP. Although the
area of highly suitable habitat has also increased among regional public lands, an
expanding bear population will require more highly suitable habitat than exists only
among public lands. Therefore, we emphasize the importance of maintaining highly
suitable habitat among timberlands, particularly in proximity to public lands. Although
future changes in land use will likely affect changes in land cover and timberland
ownership, it is critical that areas needed for meeting bear habitat management and

conservation goals be quickly identified and prioritized.

64

3.1. Introduction

Sacredness of protected areas is a result of the perception of them as exceptional
models for both biodiversity conservation and wilderness protection (Pimm and Lawton
1998; Soule and Sanjayan 1998). However, increasing human activity on private lands
near protected areas is resulting in a decrease in the area and quality of wildlife habitat
for many species that travel beyond protected area boundaries. In fact, human activity on
private land in proximity to protected areas is considered one of the greatest threats to
wildlife populations, and particularly to large mammalian species (Woodroffe and
Ginsberg 1998; Liu et al. 2001). As a result, species management on private lands is
becoming increasingly important for wildlife conservation to ensure the availability of
additional habitat when public lands alone do not adequately accommodate species.

Although private lands dominate the United States, conservation efforts have
focused primarily on public lands (Norton 2000). Approximately half of all listed species
occur exclusively on private lands, as do portions of ranges of almost all federally listed
species. Consequently, species conservation on private lands is as important as or even
more important than public land conservation (Knight 1999). However, private land use
can sometimes act as a barrier to species conservation agendas. For example, McDonald
et a1. (2001) surveyed landowners in proximity to a national wildlife refuge to determine
whether offering incentives would successfully allow refuge managers to acquire land
parcels adjacent to the existing refuge. Without incentives, <50% of proposed acquisition
parcels would be purchasable within the next 20 years and, ultimately, US Fish and
Wildlife Service goals for land acquisition would not be achieved (McDonald et al.

2001).

65

In the south-central US, timber management on private lands has helped retain
extensive forested areas that are also important as wildlife habitat (Rudis and Tansey
1995). Throughout North America, research results have suggested that particular timber
management regimes can be either beneficial or detrimental to particular species and
overall species richness (Fisher and Wilkinson 2005; Hagar et al. 2004; Anderson and
Crompton 2002). At present, as efforts are underway to recover many locally extirpated
species (Reading and Clark 1996), a challenge to wildlife researchers is determining how
land use, such as timber management, may affect species targeted for recovery.

In southeast Texas, management for recovery of the Louisiana black bear (Ursus
americanus luteolus) is heavily dependent upon habitat conservation among private lands
where timber management is the predominant land use. The Louisiana black bear’s range
has been greatly reduced since the early 19005. During the 19905, bear interest groups
and wildlife managers initiated a campaign to recover bear populations in Louisiana.
Two management plans were created, one for Louisiana (Bowker and J acobsen 1995),
and another for the bear’s collective historical range of Louisiana, southern Mississippi,
and southeastern Texas (BBCC 1997). To date, recovery efforts in Louisiana have been
successful (Van Why 2003), and support for recovery exists in Mississippi (Shropshire
1996; Bowman 1999). Although no breeding bear population exits in southeastern Texas.
the number of bear sightings within the region has increased during the past decade. This
is likely a result of transient bears from Arkansas, Oklahoma, and Louisiana that wander
into Texas. In Texas, black bears are listed as “threatened” (Bowker and J acobsen 1995)
and recent sightings have prompted the creation of a Texas black bear management plan

(TPWD 2005). The goal of the plan is to restore habitat for the purpose of reestablishing

66

black bear as a viable ecosystem component in southeastern Texas (TPWD 2005).

More than 75% of land in southeastern Texas is privately owned by timber
management companies or publicly administered by the USDA Forest Service or
National Park Service. Public lands such as Big Thicket National Preserve (BTNP),
which was established in 1974 as the nation’s first National Preserve to protect the unique
ecosystems found within the area, are potential target areas for bear population recovery
(Garner 1996; BBCC 1997; Epps 1997). However, total area and distribution of public
lands alone are not large enough for maintaining a viable black bear population. If bears
are released into BTNP, the preserve’s disjunct distribution may result in increased
movement of bears between preserve units (see Section 3.2.1). As a result, private lands
will play an important role in the success of a bear recovery. Unfortunately, in 2002 a
large number of timberlands within southeastern Texas were divested because of an
economic decision by the Louisiana-Pacific Corporation (LP), which could cause
acceleration of forested land conversion to other uses (e. g., residential development).
The purpose of this study was to determine relationships between timberlands and
potential bear habitat, and how divestment of timberland may affect area and distribution
of potential bear habitat if timberlands are converted from forest to other land uses (e. g.,
residential development). Therefore, our objectives were to: (1) quantify how area and
distribution of potential bear habitat and timberlands have changed since establishment of
BTNP, (2) quantify the relationship between highly suitable habitat and distribution of
timberlands, and how this relationship has changed over time, (3) quantify the
relationship between timberlands and areas of consistent, improved, or degraded habitat

quality over time, (4) compare changes in habitat quantity and quality over time among

67

private lands with such changes among public lands, and (5) assess implications
regarding changes in timberlands and how they may affect attainment of bear

management goals by the Texas Parks and Wildlife Department.

3.2. Methodology
3.2.]. Study area

Our study area included 12 counties in southeast Texas (Figure 3.1a). These
counties include and surround the >37,000 ha BTNP (Lat/Long. —94.36/30.38; Figure
3.1b), portions of the Sam Houston, Davy Crockett, Angelina, and Sabine National
Forests, and extensive timberlands (see Results). The southern portion of Liberty County
was excluded from analysis because of absence of timberlands and almost complete
unsuitable habitat as a result of extensive row-crop agricultural and residential land use
along the Gulf Coastal Plain. The landscape is mostly rural, but scattered with numerous
small towns and one larger city (Lufltin). Approximately 500,000 residents live within
the study area (US Census Bureau data), but mostly along the southern edge within

suburban fringes of Houston and Beaumont.

3. 2. 2. Imagery classification

Recent land cover. «Using Imagine 8.7 (Leica Geosystems GIS & Mapping, LLC,
St. Gallen, Switzerland), we derived present land cover classification from 2002
LandsatTM ETM+ (30m resolution), multispectral imagery. The study area included
portions of four 2002 LandsatTM scenes. We initially sought images from November

because of the typical lack of cloud cover in southeast Texas during late autumn.

68

However, because no cloud—free late-autumn images for all four scenes were available,
we supplemented November images with March data. To minimize impact of using
image data from different months, training sites were selected such that land cover was
easily and consistently identifiable across seasons. During June 2002 and 2003, we used
a GarminTM hand-held global positioning system (GPS) to ground-truth 255 training sites
for application of land cover data to imagery classification. Training sites were grouped
into seven general land cover categories: (1) urban-residential, (2) water, (3) sand, (4)
agriculture-grazing, (5) pine forest (both plantation and natural), (6) mixed (pine-
hardwood) forest, and (7) bottomland forest (hardwood forest within bottomland areas;
adapted from Harcombe and Callaway, 1997).

We recoded land cover data into four black bear habitat classes: highly suitable,
suitable, marginal, and unsuitable. Assignment of land cover to a specific habitat class
was based on black bear ecology for the south-central U.S. (Pelton 1982; Weaver et a1.
1990; Marchinton 1995; Larkin et al. 2004). For example, hardwood, mixed hardwood-
pine, and bottomland hardwood forests within the southeastern US contain a variety of
mast-producing species (e.g., acorns) that are often consumed by black bears (Bowker
and Jacobson 1995; BBCC 1997) and were designated as highly suitable habitat.

C learcuts and downed slash were designated as suitable habitat because early
successional vegetation often contains berry-producing species for foraging, and thick
vegetation for cover (Pelton 1982; Weaver et al. 1990). Marginal habitat was identified
as pine plantations and open sandy areas that may be traversed by bears moving between
more suitable habitat, but likely not inhabited for long periods of time. Urban and

agricultural areas were identified as unsuitable because of high probability of contact

69

with humans. Please refer to Chapter 2 for complete detailed description of habitat
classes and associated vegetative species.

Past land cover. --To determine how potential bear habitat has changed over time,
we classified imagery at approximately 10-year intervals since establishment of BTNP.
Although we attempted to be consistent between time periods, limited availability of
cloud-free imagery led us to use the following imagery data: (1) 19705: December 1974
and March 1975, (2) 19805: October and November 1986, and (3) 19905: October and
May 1997, and (4) November and March 2002, the first year of this study. Because
classification of past imagery from the 19705-19905 based on present ground-truthing
data (2002 and 2003) was impossible, we based supervised classification of 19705,
19805, and 19905 imagery on comparisons to areas of unchanged land cover from 2002
image data (Liu et al. 2001).

Analysis of land cover change. «We used Imagine’s MODELER function to
assess how bear habitat has changed over time. MODELER incorporates Spatial
Modeler Language (SML) into GIS and image processing operations. We identified all
combinations of habitat changes that could take place between the four time steps used.
With four habitat classes and four time periods, this resulted in 256 possible
combinations of habitat change. We grouped the 256 combinations of habitat change into
three categories: (1) consistent habitat, or pixels of consistent habitat class for all four
imagery time periods, (2) habitat improvement, or pixels consisting of classes that
suggested an increasing trend in the quality of habitat between the 19705 and 2002, and
( 3) habitat degradation, or pixels consisting of classes that suggested a decreasing trend in

the quality of habitat between the 19705 and 2002. For example, one possible

70

combination was that a pixel might have been classified as marginal (habitat) during the
19705, suitable during the 19805, highly suitable during the 19905, and highly suitable in
2002. The overall temporal trend for this example suggested improvement in habitat
quality over time.

Analysis of habitat among timberlands and public lands.--Using Imagine, we
georeferenced and digitized available timberland ownership maps from approximately
1970 (Acme Map Company, Tyler, Texas), 1990 (Temple-Inland, Inc.), and 2002 to
derive the distribution of timberlands across the study area. These maps did not match
time periods of imagery perfectly, but provided us an estimate of how much habitat is
related to distribution of timberlands. Similarly, we used public lands data available for
ArcView GIS 3.2 (Environmental Systems Research Institute, Inc., Redlands, California,
USA) to identify locations of public lands within the study area. We used MODELER to
select areas of specific land cover (i.e., bear habitat) that corresponded to presence of

timberlands and public lands for each respective time period.

3.3. Results
3. 3. 1. Changes in area and distribution of bear habitat and timberlands over time

Total area of highly suitable habitat across the study area has decreased overall by
approximately 150,000 ha since 1974 (Figure 3.2). However, the area of highly suitable
habitat has increased over the past decade. Suitable habitat area has remained relatively
the same, whereas marginal habitat has increased after a steep decline from the 19705 to

19805. Area of unsuitable habitat increased through the 19905, but decreased

71

dramatically between the 19905 and 2002, corresponding with increases in highly suitable
habitat.

During the 19705, timber management companies owned 696,754 ha (27%) of
lands within the study area (Figure 3.3), which were distributed mostly across the
northwestern portion of the area (Figure 3.4a). By the 19905, the distribution of
timberlands was more extensive (Figure 3.4b), and included 1,203,820 ha (47%) of the
area. A 10% decrease in timberland to 944,739 ha between the 19905 and 2002, which
mostly affected central and eastern portions of the area (Figure 3.4c), resulted from
Louisiana Pacific’s land divestment. The divestment affected approximately 25% of
timberland within each habitat class (Figure 3.5), and was distributed between units of

BTNP and along the Texas-Louisiana border (Figure 3.4d).

3. 3. 2. Relationship between timberlands and highly suitable habitat over time

In 2002 (post LP divestment) approximately one third of highly suitable (36%),
suitable (31%), and unsuitable (36%) habitat, as well as 27% of marginal habitat, found
throughout the study area was contained within timberlands (Figure 3.3). This is an
general overall increase of these three habitat classes since the 19705 when 23%, 18%,
26%, and 26% of timberlands were highly suitable, suitable, marginal, and unsuitable
habitat, respectively, but a decrease from the 19905 of 41%, 52%, 47%, and 39% for the
same classes. Therefore, during the 19705 when timberlands consisted of 27% of the
area, a relatively proportional amount of each habitat class was found among them, as

was also the case during the 19905 and 2002.

72

3.3.3. Relationship between timberlands and changes in habitat quality

Timberlands contained almost half of all land within the study area that was
consistently highly suitable habitat, and >30% of consistently highly suitable habitat
since establishment of BTNP (Table 3.1). In addition, timberlands also contained
approximately 30% of land that has either improved or degraded in habitat quality since
the 19705 (Table 3.2). Therefore, presence of timberlands has played an important role in
the maintenance of highly suitable habitat over time. The 2002 timberland divestment
involved >10% of all consistently highly suitable and marginal land across the study area,
but <10% of suitable and unsuitable area. Because a greater proportion of LP lands
improved, rather than degraded, in habitat quality over time, there is uncertainty as to
whether the 10% of highly suitable habitat directly affected by LP ’5 divestment will
remain highly suitable in the future. More land that has improved in habitat quality over
time (9%) is potentially lost as a result of divestment than land that has degraded in

habitat quality (3%) over time.

3. 3. 4. Changes in habitat area and quality among public lands

Total area of public lands within the study area was 175,070 ha, of which 37,000
ha is BTNP. Thus, post-LP investment, public lands constitute approximately 18% of the
total study area. Since establishment of BTNP, the area of highly suitable habitat among
public lands (i.e., BTNP and National Forests) has increased (Figure 3.6). In 2002, area
of suitable habitat was approximately equivalent to that in the 19705. Area of marginal

land has decreased during the past decade, whereas area of unsuitable land has remained

73

relatively consistent. In 2002, the area of highly suitable habitat was more than three

times the area of any of the other three habitat classes.

3.4. Discussion and conclusions

As humans develop areas surrounding public lands, private lands are playing an
ever-increasing role in wildlife conservation. Within the eastern US (as compared to the
western US), small reserves have presented a constant challenge to wildlife conservation
(Kulhavy and Conner 1986). Rudis and Tansey (1995) concluded that, throughout the
southeastern US, reserves alone were too small to conserve large carnivore species.

One objective of the Louisiana black bear restoration plan is to establish five
subpopulations of bears within one metapopulation, and maintain corridors for
immigration and emigration between subpopulations (BBCC 1997). BTNP and the
Middle Neches River area (which includes portions of BTNP) were identified as potential
bear release sites for subpopulation establishment in southeast Texas (Garner 1996; Epps
1997). Because home ranges and movement distances of black bears can range from
<10-100+ km2 (Pelton 1982; Weaver et al. 1990; BBCC 1997; Lariviére 2001), BTNP’s
potential as bear refuge will likely force bears traverse private lands as they move among
the preserve’s relatively small disjunct units. Even when considering National Forests
along with BTNP, public land as a percentage of total area is relatively small (<20%).

Bears are adaptable to timber management as long as forests fragmented by
harvest remain accessible and provide bears with vegetative species for foraging and
slash piles for denning (Hillman and Yow 1986; Weaver et al. 1990). For bear recovery

goals in southeast Texas, temporal change in distribution of highly suitable habitat among

74

timberlands is likely as critical as potential permanent widespread loss of forestland.
Since establishment of BTNP, area of highly suitable habitat across the region has
remained fairly constant, with a net loss of <10%. Regardless of the high concentration
of highly suitable habitat among them, because public lands make up a relatively small
portion of the study area there are likely few locations that will indefinitely be immune to
changes in land use. In fact, BTNP is relatively small compared to the National Forests.
and the effect of land use change (e. g., conversion of timberlands to residential
development) on BTNP will likely be more pronounced than on National Forests. Bear
managers may also want to explore the possibility of National Forests as potential
locations for bear population recovery.

Timberland distribution changed dramatically between the early 19705 and mid-
19905. Almost half of consistently highly suitable habitat was located among
timberlands, as well as a relatively equal proportion of land that has experienced
improvement or degradation in habitat quality. Therefore, the mere presence of
timberlands has likely been an important factor in general maintenance of high quality
bear habitat across the study area. Even after divestment by LP in 2002, >400,000 ha of
highly suitable bear habitat existed among remaining timberlands. This includes almost
90% of habitat that has remained highly suitable since establishment of BTNP. The
potential loss of >1 50,000 ha of forest is disappointing from an environmental standpoint,
but the distribution of divested lands is likely of greater significance to bear recovery and
management goals. The LP divestment affected large areas of forested lands in proximity
to BTNP, particularly between the northernmost preserve units (Figure 3.4c). Potential

loss of timberland in this area may decrease the ability of bears to move between preserve

75

units, and also between BTNP and National Forest land to the north. Therefore, we
suggest that bear managers seek to conserve bear habitat between units of BTNP. as well
as between BTNP and National Forest lands.

Private land in the northern portion of the study area will potentially be important
for bear dispersal and movement between public lands. However, maintaining forested
areas along the region’s eastern edge (Louisiana border; Figure 3.4c,d) will be even more
critical for bear movement between Louisiana and Texas, which is one of the black bear
recovery plan metapopulation objectives (BBCC 1997). Southern portions of divested
land may be prone to residential development, as this area is within a moderate
commuting distance from towns within Orange County. Timberlands along the eastern
edge of the region are dispersed among bottomland hardwood and cypress forests along
the Sabine River’s floodplain (the Sabine River forms the border between Texas and
Louisiana). If bears are released into BTNP, they must travel unimpeded through this
area to emigrate to Louisiana populations, and vice versa. As a result, we also suggest
bear habitat conservation efforts focus on divested lands along border of Texas and
Louisiana.

Timber managers in southeastern Texas have made efforts to practice harvesting
methods that promote maintenance of natural hardwood and herbaceous species for
wildlife habitat. Although this is extremely commendable from a conservation
perspective, it may not be enough to maintain bear habitat in southeast Texas over time.
Like many locations worldwide, the human population and number of households within
southeastern Texas are growing rapidly (US Census Bureau data), and uncertainty exists

over future land use decisions. In 2004, the National Parks Conservation Association

76

(N PCA) identified BTNP among “America’s Ten Most Endangered National Parks”
(NPCA 2004). Besides loss of nearby timberlands, other challenges facing regional
conservation efforts include pr0posed residential development of luxury homes, increased
petroleum extraction, and water use projects (e. g., dam construction and improvements).
In addition. planning for a Trans Texas Corridor Project includes expansion of a US
highway, which passes through the preserve, into a 366-meter wide eight-lane interstate
superhighway that would eventually connect Chicago to Brownsville (N PCA 2004).
Completion of this project would likely result in escalating pollution and development
and additional (and potentially almost impassible) fragmentation to forested areas.
Threats such as these to biodiversity conservation are a regular occurrence in proximity
to both public and private lands, and that conservation within public lands are not always
enough for meeting management goals. We advocate awareness of the difference that
conservation management on private lands can make. and suggest incorporation of

private land initiatives into conservation planning.

77

Table 3.1. Area of land (ha; percent total area) that has remained consistently within a
single habitat class since establishment of BTNP, and area of land (ha; percent of total
area) that was affected by divestment of land by Louisiana-Pacific (LP).

 

 

Habitat class Whole area Timberland 2002 LP divestment
Highly suitable 257,965 122,353 (47%) 29,120 (11%)
Suitable 9,168 2,894 (32%) 560 (6%)
Marginal 11,122 6,115 (55%) 1,528 (14%)
Unsuitable 55,241 3,089 (56%) 624 (1%)

 

78

Table 3.2. Area of land (ha; percent total area) that has improved or degraded in habitat
quality since establishment of BTNP, and area of land (ha; percent of total area) that was
affected by divestment of land by Louisiana-Pacific (LP).

 

 

Land area Improvement Degradation
Whole area 168,885 360,485
Timberlands 50,674 (30%) 102,716 (28%)
LP divested lands 15,535 (9%) 11,407 (3%)

 

79

Figure 3.1. (a) The 12-county study area in southeastern Texas. The southern portion of
Liberty County was excluded from analysis because of absence of timberlands and
almost complete unsuitable habitat as a result of extensive agricultural and residential
land use. (b) USDA proclamation boundaries for the Sam Houston, Davy Crockett,
Angelina, and Sabine National Forests (striped; clockwise from bottom left) and Big
Thicket National Preserve (dark shaded).

80

 

 

 

 

 

 

W mg “if” in
1

1111111111 fl.

l——l
25km

.011

160077 7 , 777 7a 7

   

Area (1000 ha)

 

 

1400 77 7 , r , ’ , "I---
12119705
1200 7 7 , t * t , * *'*7'I19805*7’7
.l19905
10007 7 7 7 7 7 7 7 7 7 777-30037777
8007 7 7 777 77777
60077 7777777777
400 77
200 7
O _,
Marginal Unsuitable
Habitatclass

Figure 3.2. Area (1,000 ha) of highly suitable, suitable, marginal, and unsuitable bear
habitat at 10-year intervals since establishment of BTNP.

82

 

1400

 

 

 

1200 7
1000 - I 0me area if
,3 El Timberland 19705.
‘3 800 ITimberhnd19905
O
93 -'_T 111336992992
E 600 ~
<
400 7
200 7
O _
Highly suitable Suitable Marginal Unsuitable
Habitat chss

Figure 3.3. Area (1,000 ha) of land within each habitat class across the study area
compared to land contained within timberlands since establishment of BTNP.

83

Figure 3.4. Changes in distribution of timberlands (red) across the study area, since
establishment of BTNP, based on available timberland ownership data. Also illustrated
are public lands including Big Thicket National Preserve (blue) and USDA National
Forest proclamation boundaries (green). (a) During the late 19605, before establishment
of BTNP, timberlands were concentrated within the northwestern portion of the study
area. (b) During the 19905, timberlands became more evenly distributed across the study
area. (c) Orange areas illustrate land divested by Louisiana—Pacific in 2002, which are
largely concentrated in the central and eastern portions of the study area. ((1) Remaining
timberland in 2002 after divestment of lands owned by Louisiana-Pacific. County
boundaries are designated by black lines.

84

4‘7

.. .MJ ..
I ..fla”.

85

 

Figure 3.4

Figure 3.4 (Cont’d.)

(C)

 

86

 

\l
O
O

O\
O
O

I’Loui’sblapacilic landdivestEd‘
I 3“er mm 2,097: L

b) 4} UI
O O O
O O 0

Area (1000 ha)

N
O
O

100 7

 

 

 

Highly suitable Suitable Margiml Ursuitable
Habitat class

Figure 3.5. Area (1,000 ha) of land divested by Louisiana Pacific, Inc., in 2002. Among
timberlands, divestment affects approximately 20-25% of land within each habitat class.

87

140

 

     
   

 

 

 

 

120 777777 7 7 7 777 77 777777777

100, 7, , , lil’rlblichrldilétosgg_i ,
E IPublic hnd19905
8 so 7 , 77777 7 7.IPublichrld2002 77777777
2 , , , ,,,,,
§ 60 77'
<2

40 77

20 7 7

0
Highly suitable Suitable Margiml Unsuitable
Habitatclass

Figure 3.6. Area (1,000 ha) of public land within each habitat class across the study area
since establishment of BTNP.

88

CHAPTER 4

ATTITUDES AFFECTING SUPPORT FOR A PROPOSED BLACK BEAR

REINTRODUCTION IN EAST TEXAS

In collaboration with

Angela G. Mertig, Jianguo Liu, and Nathan Garner

89

Abstract

Socioeconomic knowledge about local residents is critical for survival of
reintroduced species in human-occupied locations. Our study provided a unique
opportunity to assess public attitudes prior to an anticipated black bear reintroduction.
We surveyed 3,000 residents of East Texas (40% response rate) to evaluate attitudes
toward 1) black bears, 2) wildlife managers, and 3) a potential increase in the black bear
population. Positive attitudes toward all three were related to more frequent participation
in passive-appreciative activities related to wildlife (e. g., observing or reading about
wildlife) and greater knowledge about bears. Men, younger residents, and individuals
who own fewer acres of land had more favorable attitudes toward both black bears and
increasing the bear population. Respondents who have lived in the area for a shorter
amount of time also had more positive attitudes about bears. Urban residents and
individuals who own fewer acres were more confident in the ability of wildlife managers
to manage bears. Lack of knowledge about bears was the most commonly stated reason
why respondents were unsure about increasing the bear population. Distributing
information about bear ecology and behavior may foster additional support for a
reintroduction. Just as important, managers must understand particular reasons for
resident support or opposition, and whether specific incentive may result in tolerance of

bear presence regardless of attitudes.

90

4.1. Introduction

While the amount of wildlife habitat decreases, increasing recognition of
carnivores as valuable members of the ecological community has resulted in efforts to
recover populations of locally extirpated species (Reading and Clark 1996, Sillero-Zubiri
and Laurenson 2001). Recovery is often difficult for large species, such as black bear
(Ursus americanus), that require large areas of habitat that may no longer exist as a result
of human activity, or are under private ownership (Pelton 1986, Meffe and Carroll 1997,
Woodroffe 2001a). For species reintroduction programs in human-occupied locations,
socioeconomic knowledge about humans who reside in or make use of an area is vital for
ensuring long-term survival of a resident species. Although wildlife management has
traditionally focused on ecological dynamics of extant wildlife populations, substantially
more attention has recently been placed on the needs and wants of humans who may be
affected by management decisions (Kleiman 1989, Riley et al. 2002).

Reintroduction—related management has evolved as a result of past policy failures
(Kellert 1991, Reading and Kellert 1993). Large species with negatively perceived
consequences (e.g., danger to humans) have provoked negative feelings in humans for
several reasons, such as fear, unfamiliarity, negative personal experiences (e. g., nuisance
activity such as pilfering through garbage), or negative hearsay (Kellert et al. 1996).
Consequently, maintaining a socially suitable environment (i.e., human acceptance of a
species) is often more challenging than maintaining a biologically suitable environment
for a species.

In North America, negative attitudes and opinions are often associated with

carnivores, such as black bear and wolves (Canis lupus). The process of reintroducing

91

wolves to Yellowstone is probably the most publicly visible and well-known recent
reintroduction program. Knowledge about wolves, land use association (e. g., rancher,
wildlife activist), age, and location of residence (i.e., rural or urban) were factors that
influenced attitudes toward wolf reintroduction, although this information was not
directly used to determine reintroduction feasibility (Bath and Buchanan 1989). Beyond
Yellowstone, local ranchers were greatly influential in reintroduction planning for
Mexican gray wolves (Canis lupus baileyi) in Arizona (Schoenecker and Shaw 1997). In
Michigan, although more knowledgeable about wolves than Lower Peninsula residents,
Upper Peninsula ranchers and farmers were more likely to oppose wolf restoration
(Kellert 1991). More-educated respondents, women, self-expressed animal lovers, urban
residents, and residents with most knowledge about wolves were more likely to support a
wolf reintroduction in New Brunswick (Lohr et al. 1996). These results implied that
particular demographic and economic variables influence attitudes toward, and opinions
about, wolf restoration. Such findings may be applied to reintroductions of other species.

Like wolves, black bears have often garnered a wide array of public reactions. In
Arkansas, no public input was sought regarding a black bear reintroduction during the
19605 because the Arkansas Game and Fish Commission perceived that the general
public would not support the program (Smith and Clark 1994). Although black bear
population recovery in Arkansas was biologically successful, release of bears without
public input would be unacceptable by today’s wildlife management standards, of which
public accountability is an important component (Smith and Clark 1994). Several studies
have been conducted to address management for potential black bear population

increases (Decker et al. 1981, Decker et al. 1983, Decker and O'Pezio 1989, Bowman et

92

al. 2001, Peyton et al. 2001; Van Why and Chamberlain 2003; Bowman et al. 2004) and
biological feasibility of population recovery (van Manen 1991, Garner 1996, Epps 1997).
Fewer studies have addressed social feasibility of black bear reintroductions.

Black bear reintroduction efforts, where evaluated, have been met with mixed
support, but overall attitudes toward black bears have been generally positive. Kellert
(1994) reported that North American positively perceived bears as intelligent and
aesthetically appealing. In a survey of visitors to Great Smoky Mountain National Park,
most respondents indicated that bears were an important component of the park even if a
negative encounter was experienced (e. g., property damage, Pelton et al. 1976). In both
Massachusetts (J onker et al. 1998) and Mississippi (Bowman et al. 2001), some bear
damage to agricultural products was considered tolerable. In fact, Massachusetts
residents believed that presence of bears held aesthetic, ecological, and economic value
(.lonker et a1. 1998). In New York, approximately one-third to one-half of survey
respondents wanted to see a black bear on their property (Decker et a1. 1981). Any
successful reintroduction will depend upon social acceptance among local residents

(Clark et al. 2002).

4.1. 1. Return ofthe Louisiana black bear

The Louisiana black bear (Ursus americanus luteolus) is one of the sixteen
subspecies of American black bear (Pelton 1982). Its historical range included Louisiana,
East Texas, and southern Mississippi. Overharvest and habitat destruction led to the near
demise of the subspecies by the mid-19505 (BBCC 1997). Bear populations in southeast

Texas were extirpated by the early 19005. At the end of the 19805, only two small

93

populations of the Louisiana subspecies remained, both in eastern Louisiana. During the
19905, extensive public outreach led to recovery efforts in Louisiana and creation of
recovery plans throughout the subspecies’ historical range (Bowker and Jacobson 1995,
BBCC 1997). Recovery in Louisiana has been deemed successful thus far (Van Why
2003, Benson 2005) and feasibility analyses have been completed in Mississippi
(Shropshire 1996, Bowman 1999). Our study, coupled with recent attitude surveys in
Louisiana and Mississippi, enables comparison of attitudes toward the Louisiana black
bear’s recovery across its entire historical range.

In East Texas, the number of black bear sightings has increased during the past
decade. Because there is no known breeding population present, it is likely that these
sightings are transients from Arkansas, Oklahoma, and Louisiana. This increase in
sightings has prompted the creation of a black bear conservation and management plan
for Texas. Objectives of the plan for the next 10 years are public coordination,
communication, outreach/information dissemination, habitat management, and research
(TPWD 2005). The ultimate goal, through partnerships with other resource management
organizations and members of the general public, is to restore habitat for the purpose of
reestablishing black bear as a viable ecosystem component in East Texas (TPWD 2005).

This study represents a unique opportunity to assess public attitudes toward black
bears prior to an anticipated reintroduction. This research reports social survey
information regarding public attitudes toward the Louisiana black bear in the southern
portion of East Texas. Our objectives were to (1) identify public attitudes toward black

bears and a potential reintroduction. (2) determine what factors influence residents’

94

attitudes toward recovery of black bear in the area, and (3) compare this information with

attitude studies from the remainder of the Louisiana black bear’s historical range.

4.2. Methods and measurements
4.2.]. Study area

Our study area consisted of 12 counties (Area = 25,371.63 kmz) within the
southern portion of East Texas (Figure 4.1). Approximately 500,000 residents live within
the study area, and approximately 368,000 of residents were adults in 2000 (218 years of
age; US. Census Bureau data). Much of the area is considered rural, but dotted with
numerous small towns and one larger community (the city of Lufldn, with approximately
50,000 people). More than 75% of the land is privately managed for timber or owned by
the Federal government (Big Thicket National Preserve and the Davy Crockett, Sabine,
Angelina, and Sam Houston National Forests). The more-densely populated southern

edge of the area includes suburban development related to Houston and Beaumont.

4. 2. 2. Survey design

Based on the distribution of the human population within the area, as determined
by data from the 2000 Census, we divided the study area into three strata: 1) rural =
residents residing in rural areas, as determined by comparing population density between
zip codes using ArcView GIS 3.2 (Environmental Systems Research Institute, Inc.,
Redlands, California, USA), 2) urban = residents residing in towns with >2,500 people
and the City of Lufkin, and 3) south = residents residing in suburban areas along the

southern portion of the study area. In January 2004, we mailed a questionnaire to 3,000

95

residents among the three strata: 2,000 rural, 600 urban, and 400 south (Appendix). Our
target sample of 3,000 residents was based on funding, and the number of residents
sampled per stratum was based on the proportion of the study area’s population within
each stratum. Assuming that rural residents will likely have the greatest probability of
contact with bears (based on black bear ecology; Pelton 1982), we adjusted our sampling
proportions in order to over sample, and therefore adequately represent, the rural stratum,
which also contained the fewest residents. Name and address information were
purchased from Survey Sampling, Inc. (Fairfield, CT), a company that assembles
representative survey samples from telephone directory listings. A modified version of
the “Tailored Design Method” (Dillman 2000), which uses multiple contacts and careful
attention to survey details to increase response rate, was applied to survey design and

implementation.

4. 2. 3. Dependent variables

We used exploratory factor analysis (principal components analysis; Babbie
1990) to define two dependent variables based on a series of questions related to black
bears (Table 4.1): 1) attitudes toward black bears, and 2) attitudes toward wildlife
managers. A third variable, attitude toward population change, measured attitudes related
to whether the black bear population in East Texas should increase (i.e., more bears in the
area), decrease (i.e, no bears in the area), or remain the same (i.e., allow transients to
move through). Questions related to all three dependent variables were asked in Likert-
scale format, a five-response scale ranging from strongly agree to strongly disagree (5 =

strongly agree, 4 = agree, 3 = unsure, 2 = disagree, 1 = strongly disagree; Babbie 1990).

96

Each of these variables reflected important elements of social acceptance of black bears
in East Texas and helped determine the potential for a socially successful reintroduction
of black bears.

Attitudes toward black bears. Seven statements, related to respondent perceptions
of and emotional feelings about black bears, grouped together in factor analysis (Table
1). These statements included whether the existence of black bears 1) is a sign of a
healthy environment and 2) would increase their (resident’s) quality of life in East Texas
as well as 3) near their home; 4) whether bears have the right to exist wherever they
occur; whether a resident 5) felt personally at risk if black bears exist and 6) was afraid of
black bears; and 7) their opinion as to whether black bears commonly harm humans. To
accommodate missing data, respondents who answered <75% of questions related to
attitudes toward black bears were eliminated from analysis. If a respondent answered
275% of questions, any missing data were assigned the average value across all
respondents for that item. Items were coded so that larger values reflected greater
support for black bears and then summed. A test for internal consistency (or reliability)
among responses resulted in a Cronbach’s alpha score of 0.86 (Cortina 1993).

Attitudes toward wildlife managers. Two questions relating to management
loaded on an additional factor (Table 1). First, we asked whether respondents agreed
with the statement that wildlife managers know how to manage bears. Second, we asked
whether participants agreed that wildlife managers understand landowner concerns about
bears. The management variable consisted of a summation of these two items. Larger
numbers reflected greater confidence in managers. Pearson’s correlation between the two

items was 0.691. Any cases with missing data were deleted.

97

Attitudes toward population change. Participants were asked to respond to the
statement, “The black bear population in East Texas should be increased.” Larger values
derived from Likert scale format reflected greater support for a black bear population

increase.

4. 2. 4. Independent variables

We selected 17 independent variables for our analysis (Table 4.2). Although we
measured race/ethnicity, it was not included because 95% of the respondents were
white/Caucasian. This underrepresented members of racial and ethnic minority
populations, which constituted 23.7% of the total population of the area (US. Census
Bureau data). Based on findings and reasoning from similar studies, we hypothesized
that males, younger respondents, more-educated respondents, those with higher incomes,
and respondents who were more knowledgeable about bears would be more likely to
support black bears and an increased bear population. We also expected that members of
urban communities, people who have seen bears in the wild, newer residents to the area,
members of a wildlife-related organization, and residents who participate more frequently
in activities related to wildlife would have more positive attitudes toward bears wildlife
managers, as well as be more supportive of increasing the bear population. Finally, we
hypothesized residents with children, pet owners, large-tract landowners, members of the
rural stratum, and individuals who tend livestock would be less amicable toward bears,
wildlife managers, and increasing the bear population. Four of the independent variables

(described below) are composites of separate items from our survey.

98

Activities. We asked respondents to report participation in 16 activities that may
put them in contact with bears and other wildlife (Table 4.3). Results from factor
analysis suggested data reduction to three activity categories: utilitarian, passive
appreciative, and work-related. Utilitarian activities included camping, boating, all-
terrain vehicle use, hunting large game, hunting small game, and fishing (Cronbach’s
alpha = 0.84). Passive appreciative activities included reading about wildlife, watching
TV shows or movies about wildlife, and observing wildlife (Cronbach’s alpha = 0.79).
Work-related activities included working on a farm or for the timber industry (r = 0.31).
Hiking, running, biking, canoeing, and working for the oil industry did not factor into one
of the above categories and were eliminated from further analysis. For each set of
activities retained for analysis, respondents who answered <75% of the related questions
did not receive an activity score. If a respondent answered 275% of the related questions.
any missing data were assigned the average value for that item across all respondents.
For each set of activities, a scale score was derived by summing up responses for each
item. with larger values reflecting greater levels of participation in the activities.

Knowledge. Respondents were asked to indicate (yes or no) for each of the
following five factual statements about black bears in the region: 1) until the early 19005.
eastern Texas contained a large population of black bears, 2) the number of black bear
sightings in eastern Texas has increased during the past decade, 3) black bear populations
are increasing in size in Arkansas, Louisiana, and Oklahoma, 4) black bears in Texas are
protected by both federal and state legislation, 5) black bears exist throughout most of the
United States and North America, and 6) black bears are mainly vegetarians. A score of

1 was given for each “yes” indicated by the respondent, which assumed indication that

99

the respondent was knowledgeable about the particular fact. A score of 0 was given for
each “no.” Scores were then summed to create an overall knowledge score for each
individual. Respondents who answered <75% of the related questions did not receive a
knowledge score. If a respondent answered 275% of the related questions, any missing
data were assigned a 0. A knowledge score was derived by summing up responses for

each item.

4. 2. 5. Non-response follow-up

Because our response rate was <50%, we completed a non-response follow-up for
all individuals (n = 1,600) within the survey sample who did not return a survey minus
those who indicated that they did not wish to participate as well as those for which we
had incorrect addresses. We asked 10 questions, similar to those in the actual survey, in
an effort to determine whether there was a significant difference in responses between

survey respondents and individuals who completed the non-response questionnaire.

4. 2. 6. Statistical analysis

We used SPSS 12.0 for Windows software to complete statistical analyses (SPSS,
Inc., 2003). Because we oversampled rural residents, we applied weights to descriptive
(univariate) analyses to more accurately represent the entire area; sample weights were
not used for bivariate or multivariate analyses (Babbie 1990). For bivariate analyses
(statistical tests comparing two variables), one-way ANOVA (F test) and Pearson’s r
were used to compare sample means and test relationships between variables,

respectively (Babbie 1990, Sokal and Rohlf 1995). Multivariate analyses (tests using

100

three or more variables) were completed using ordinary least squares regression for each
dependent variable (Babble 1990, Sokal and Rohlf 1995). All alpha values were defined
at the 95% confidence interval. After accounting for multiple comparisons in bivariate
analyses (17 tests per dependent variable) with a Bonferroni correction, the P values were

considered significant at the level of 0.003.

4.3. Results

The overall response rate was 40% (n = 1,006) after removing incorrect addresses
and those considered ineligible to respond (see also Morzillo et al. 2005). For each
stratum, the response rate was 41%, 35%, and 40% from rural, urban, and south areas,
respectively. Overall results from the survey did not differ significantly from those for
the non-response follow-up (n = 163). As indicated by the non-response follow-up
survey, the most common reasons for not completing the original survey, as indicated by
respondents, respondents had little or no knowledge about black bears (45.3%) or did not
like answering surveys (23.7%).

Detailed results of univariate analyses of independent variables (Table 4.2)
suggested that more than half of all respondents resided in rural areas or small towns.
owned pets, were male, and had a household income of _>_$40,000. There was, on
average, less than one child per household (SD = 1.03), and the average age of
respondents was >50 years old (SD = 15.11). Respondents had lived in East Texas, on
average, for >38 years (SD = 19.25). Eleven percent of respondents indicated that they
belong to a wildlife-related organization. Respondents indicated a relatively high level of

participation in utilitarian and passive-appreciative activities, and low participation in

101

work-related activities. Respondents own 22 acres of land, on average (SD = 84. 14), and
14% of landowners indicated that they tend livestock on their land. Twenty—three percent
of respondents indicated that they have seen a black bear in the wild. Twenty-eight
percent of respondents had high school diplomas as their highest level of formal
education and another quarter (26%) had at least a college degree. Respondents averaged
a knowledge score of 2.48 (SD = 1.77) questions answered correctly out of a possible six
questions.

For descriptive results of dependent variables (Table 4.4), attitudes toward bears
were generally positive, but many respondents were unsure whether the presence of bears
would increase their overall quality of life (46%). Disagreement existed as to whether
bears commonly harm humans (mean = 3.72, SD = 0.83), and many respondents (28%)
indicated that they are afraid of bears. Although 41% of respondents were unsure, only
18% believed that the black bear population in East Texas should not be increased.

Results of bivariate analyses (Table 4.5), after Bonferroni adjustment, suggested
that respondents with significantly more positive feelings toward bears included males,
those with greater household incomes, more frequent participants in utilitarian and/or
passive-appreciative activities, newer residents to the area, those more knowledgeable
about bears, and those who have seen a black bear in the wild. More frequent
participants in passive-appreciative activities and those more knowledgeable about bears
were more likely to agree that wildlife managers know how to manage bears and
understand landowner concerns. Respondents with a greater number of children <18
years of age, males, younger respondents, those with higher incomes, members of

wildlife-related organizations, more frequent participants in utilitarian or passive-

102

appreciative activities, those who were more knowledgeable about bears, and those who
have seen a black bear in the wild were more supportive of increasing the bear
population.

Multivariate analyses yielded a regression model for each dependent variable
(Table 4.6). Those with significantly more positive attitudes toward black bears included
men, younger respondents. more frequent participants in passive-appreciative activities.
newer residents to the area, those more knowledgeable about bears, and those who have
seen a bear in the wild. More positive attitudes toward management were significantly
more likely among respondents who reside in more-urban communities, participate more
often in passive-appreciative activities, own fewer acres of land, were more
knowledgeable about bears, and reside within the urban stratum. Attitudes supporting a
bear population increase were significantly more likely among men, younger
respondents, more frequent participants in passive-appreciative activities, those who own

fewer acres of land, and those more knowledgeable about bears.

4.4. Discussion

Similar to results from Decker et al. (1981) and Kellert (1994), our respondents
indicated generally positive attitudes toward bears. This is also consistent with Buck and
Brown (2002), Kellert (1994), and Clark et al. (2002) in that there was general support
for an increase in the black bear population size. In fact, only six percent of our
respondents indicated that black bears should not exist at all in East Texas (Morzillo,
unpublished data). This is a much lower percentage than in Mississippi where 16.8% and

51% of the general public and landowners, respectively, opposed an increase in the black

103

bear population (Shropshire 1996, Bowman 1999). Support for a bear population
increase in our study was much greater than that for wolves in Arizona (Schoenecker and
Shaw 1997). However, support for wolves in Yellowstone was high among respondents
excluding those respondents who were members of the Stock Growers Association (Bath
1989). Farmers were the only respondent group with unfavorable attitudes toward a wolf
reintroduction in Minnesota (Kellert, 1985a). Although livestock owners in our study
had significant positive attitudes toward bears before Bonferroni adjustment, further
analysis is necessary to determine whether particular user groups are supportive of bear
population increases.

In general, our regression models were relatively successful at explaining the
variance of our dependent variables (Table 4.6). The lower R—squared value for
management attitudes is undoubtedly related to the initial lower variation in this measure.
We assumed that rural residents would have a greater chance of contact with bears and
may be less supportive of a bear population increase. With all else constant, our results
suggested that more-urban communities were more confident in wildlife managers, but
differences among communities in support for a bear population increase across different
communities were not significant. Lohr et a1. (1996) found that residents in larger
communities typically had more positive attitudes toward wolves in Minnesota, as did
Bowman (1999) with bears in Mississippi. However, Peyton et al. (2001) indicated that
residents in more populated areas of Michigan preferred limited bear presence. The
demographic characteristics of our study area are changing as many residents of the area
are originally from other locations throughout the United States (Morzillo unpublished

data). In addition, many individuals are moving out of nearby cities and into more-rural

104

areas for a better quality of life (e. g., cheaper housing). The mix of urban and rural
environmental attitudes may balance regional positive and negative attitudes toward bears
and increasing the population, which may be supported by the fact that in our survey
newer residents to the area and younger respondents were more supportive of increasing
the bear population. Similarly, Schoenecker and Shaw (1997) reported that the average
residency for people with pro-wolf attitudes in Arizona was 19 years, and 24 years for
those with anti-wolf attitudes. Past research suggested generally positive local attitudes
toward wildlife managers (Duda and Brown 2001), and our results suggested no
difference.

Rural landowners may be concerned that the presence of an endangered species
will result in government restrictions on land use (Kellert 1991, Erick and Brown 2002).
In our study, although opposite from Bowman (1999), individuals who owned more land
had more-negative attitudes toward both black bear and resource managers. Open-ended
comments on our survey revealed only four respondents who felt “government is
intrusive,” or “agencies can’t manage” as reasons for not supporting an increase in bear
population size. Considering a generally positive relationship between landowners and
wildlife managers (Duda and Brown 2001), there are likely additional reasons why large-
area landowners have negative attitudes toward both bears and managers, which should
be sought in future research.

Respondents in households with a greater number of children <18 years of age
were more supportive of increasing the bear population, but only in bivariate analysis.
While we suspected that respondents might be concerned about safety of children, only

12 respondents indicated (elsewhere on the survey) that they feared for the general safety

105

of children. However, several respondents indicated that they would like their children to
have the possibility of seeing bears, or that their children were very interested in wildlife.
Two older respondents indicated that they feared for their grandchildren’s safety, but it is
possible that this concern is related to older respondents’ more-negative attitudes toward
bears and a bear population increase as compared to younger respondents.

As in past studies (e.g., Kellert 1994, Bowman, 1999), bivariate results suggested
that respondents with greater household incomes were more likely to have more positive
attitudes toward black bears and support an increase in black bear population size, but
these relationships did not hold true when holding all other variables constant. Income
was not a significant variable in Louisiana bear studies either (Shropshire 1996, Bowman
et al. 2004), and, in our study, insignificance of this variable may be related to
insignificance of other variables such as education (assuming that more educated
individuals have greater incomes). We were rather surprised to find that level of
education was not a significant predictor in any analysis, as results from many other
studies have indicated higher levels of education directly relate to greater support for
carnivore conservation and environmental protection in general (Kellert 1985b, Kellert
1994. Lohr et al. 1996, Bowman 1999).

Respondents were knowledgeable about the black bear’s past and present status in
Texas, but less aware of black bear population dynamics within the greater region. “I
don’t know anything about bears” was the most cited reason (n = 79) for indicating both
an unsure response related to a bear population increase, and for survey non-response
(45%), as indicated on the non-response follow-up. Because a completed East Texas

black bear plan was not yet available to the public at the time of this survey, we do not

106

know whether many respondents were aware of its preparation. However, Van Why and
Chamberlain (2003) reported that <60% of hunters were aware of a recovery plan in
Louisiana, and we estimate that knowledge of preparation of an East Texas bear recovery
plan was much lower. It would be interesting to compare our results to attitudes after
exposure to and marketing of the East Texas plan to see if attitudes change if respondents
become aware that the beginning stages of planning for a bear population recovery are
underway. Regardless, greater knowledge about bears was significantly positively
related to all three dependent variables in both bivariate and multivariate analyses,
suggesting that a greater understanding of the species by residents may be an important
ingredient of recovery efforts.

Respondents who had seen a bear in the wild were more likely to have positive
attitudes toward bears, which suggests that personal familiarity with the species fosters
positive attitudes toward it. Consequently, seeing a bear in the wild may cultivate an
observer’s desire to learn more about the species as a result of the observation
experience. Shropshire (1996) suggested that support for increasing the Mississippi bear
population was positively related to both knowledge about bears and whether landowner
respondents had seen a bear. However, the positive relationship between knowledge and
whether a respondent had seen a bear did not told true for the general public. Similarly,
attitudes toward and knowledge about wolves have been important predictor variables for
overall support for wolf restoration (Bath 1989, Buck and Brown 2002), even though they
may not guarantee support for reintroduction (Lohr et al., 1996). Even 50, greater
knowledge about black bears is an important element in gaining support among the

general public and residents of the area targeted for reintroduction efforts. Our results

107

highlight the importance of information dissemination and outreach from natural resource
agencies.

More frequent participation in passive-appreciative activities was significantly
related to all three dependent variables in all analyses, which suggests that participants in
these activities were more knowledgeable about and interested in the presence of wildlife
for their enjoyment. In fact, there was a significant and positive relationship between
knowledge and participation in passive-appreciative activities related to wildlife (F =
34.472, df = 6, 925, P <0.001). However, when controlling for all other variables
participation in utilitarian activities was not significant. This may suggest that
respondents participate in utilitarian activities for reasons other than those with a wildlife
focus such as hunting, such as camping to spend time with family. In other studies,
participation in activities related to wildlife was a significant factor related to attitudes
toward predators in general (Kellert 1985b), and wolves in particular (Lohr et al., 1996).
In fact, “self-professed lover of animals and nature” was a reason for supporting a wolf
reintroduction in Arizona (Schoenecker and Shaw 1997). More than 78% of sportsmen
in Louisiana supported an increase in bear populations (Van Why 2003). Although some
people may not desire the utilitarian use of a species, mere chance to observe a species
may bring them personal fulfillment.

In our study, members of wildlife-related organizations were more likely to have
positive attitudes toward a bear population increase, but only in bivariate analysis. One
reason for an insignificance of organization membership at the multivariate level may be
an underlying relationship between membership and income. Most wildlife-related

organizations require annual membership dues, which may unintentionally discriminate

108

against the ability of lower-income respondents to join such organizations. In fact, a
significant positive correlation existed between organization membership and income (r
= 0.126, P < 0.001 ). The most frequently listed organizations of membership included
Ducks Unlimited, the National Rifle Association and the National Audubon Society. As
indicated within their mission statements (on the organization’s websites), Ducks
Unlimited and the National Audubon Society are dedicated to the conservation of species
and ecosystems. Although the primary mission of the National Rifle Association is to
protect individuals’ rights to buy, own, and use firearms, the organization is also involved
in issues that are linked to firearms, which include hunting and wildlife conservation.

In social science literature related to wildlife, it is common for the respondent
pool to consist of a majority of males, and for attitudes to differ between the sexes.
Shropshire (1996) reported that 91.8% of responding landowners were male. Our results
were less lopsided in that 71.9% of respondents were males, as were 70% of the non-
response follow-up participants. This is likely influenced by the fact that we acquired our
respondents’ names and addresses from telephone listings, on which males are more
likely to be listed than females (Koval and Mertig 2004). Even if a female’s name was
listed, she may give the survey to a household male if the survey topic is of greater
interest to him. Our analyses suggested significant differences between the sexes in
respect to attitudes toward black bears and a population increase, with men being more
supportive than women. This was consistent with Kellert (1985c), but opposite of Lohr
(1996) for wolves and Peyton et al. (2001) for bears. Shropshire (1996) indicated no
significant differences between men and women as to whether the bear population should

be increased. Our results may be attributed to significantly greater male participation and

109

interest in hunting bears, as indicated elsewhere on the survey (Morzillo, unpublished
data). Consistent with Kellert (1985c, 1991) and Bowman (1999), males and respondents
who hunt had higher knowledge scores than females and non-hunters (Morzillo,
unpublished data). Concerns about children or belief that bears are dangerous animals
may be reasons for female respondents’ less positive attitudes toward bears. However, in
our study, there was no significant difference in attitudes toward black bears between
females in households with and without members <18 years old (F = 3.54, df = 1,93, P =
0.630). Regardless, future research in this study area must include more females.

Younger respondents were more likely to have positive attitudes toward bears and
support a bear population increase. Older respondents may be less supportive because of
safety concerns or resistance to change, which is consistent with results of surveys about
other carnivores (Kellert 1985b, Bath 1989, Kellert 1991). Bowman’s (1999) results
suggested greatest support for a bear reintroduction among respondents 25-49 years of
age. Several of our older respondents noted, in the general comments section of the
survey, that they remember their elders speaking of negative encounters with bears and
bears killing livestock. Such tales may have instilled a negative perception about bears
among the current elders if bears in the past were regarded as nuisances or dangerous
(Reading and Kellert 1993).

Concerns about livestock safety are often a reason for opposition toward
carnivores. In fact, this was the most-cited reason for opposition toward a wolf
reintroduction program in Arizona (65% of respondents who opposed reintroduction;
Schoenecker and Shaw 1997). Bath (1987) found similar results related to the wolf

reintroduction in Yellowstone, as did Enck and Brown (2002) in the Adirondacks.

110

Surprisingly, in our study, respondents who tended livestock held more positive attitudes
toward black bears, but only for pre-Bonferroni adjusted results at the bivariate level.
Very few respondents (n = 5) noted concern of livestock safety as a reason why they
opposed a black bear reintroduction. This simply may be because bears have seldom
banned livestock in East Texas, and respondents may recognize this. Residents who
tended livestock were also more likely to be hunters (Morzillo, unpublished data). But
results did not suggest a greater interest in hunting bears among hunters who do versus do
not tend livestock. Future research should focus in greater detail on specific stakeholder

groups.

4.5. Conclusion and implications

Little information is available about public attitudes toward reintroduction of
large mammals, especially for species often perceived as threatening. The information
that does exist with regard to black bears appears to imply generally positive feelings
toward bears and public acceptance of black bear reintroductions, with more likely
support among younger, more educated, wealthier, and urban, respondents, as well as
respondents who have more recently relocated to the area. In addition, residents who
have seen a black bear in the wild, are members of wildlife-related organizations, and
participate in activities related to wildlife also have been found to hold more positive
attitudes about black bears and increasing the bear population size. Many of our results
were consistent with past studies. We conclude that our respondents held generally
positive feelings toward black bears, and broad (but not all-inclusive) support exists for

increasing the bear population size. In particular, greater knowledge about bears and

111

more frequent participation in passive-appreciative activities were consistently and
positively related to support for black bears and an increase in their population. For other
factors that influenced attitudes, some differences existed between bivariate and
multivariate analyses; i.e.. controlling for other variables modified the original bivariate
relationship. Although managers must be aware of these differences, further analysis of
these factors is certainly warranted, as the fact that their impact varies in the presence of
other variables may make it more difficult for managers to precisely determine the
individual relative importance of each factor. Overall, a key to improving support for
black bears in East Texas, just as in Louisiana and Mississippi, is enhancing the
knowledge base (about bears) of residents through expanded outreach efforts and
encouragement of outdoor activities that promote appreciation of wildlife.

In Louisiana, where bear restoration is already taking place in various locations
across the state, survey data about attitudes toward black bears are limited to sportsmen.
Although these individuals comprise an important stakeholder group, there are other
groups of people whose voices need to be heard, if only to know how they might react if
they encounter a bear. Survey data that include the general population of Louisiana
would be useful to bear recovery efforts in Texas. Such information could aid public
outreach efforts in all three states (Texas, Louisiana, and Mississippi). Further, it is
advisable that managers in all three states cooperate in efforts to collect consistent
information from residents. This will improve the ability of wildlife managers to manage
effectively at the geographic scale of the Louisiana black bear’s range.

Survey research from Mississippi showed high support for increasing the black

bear population size and for enactment of land use restrictions in an effort to insure bear

112

protection. We do not know yet whether residents in East Texas will support land use
restrictions. Based on the information discussed above there is good cause for optimism

that the Louisiana black bear may someday be accepted throughout its historical range.

113

Table 4.1. Results of factor analysis (principal components analysis with varimax
rotation) used to define dependent variables2| for analysis of attitudes toward black bears

(N = 982) in East Texas.

 

Factor loadings for dependent variables

(Percent variance explained)

 

 

Attitudes toward Attitudes toward Otherb
black bears wildlife managers

Statement from survey (44.2%) (14.0%) (1 1.1%)

The presence of black bears is a sign 0.757 0.112 0.082
of a healthy environment

Black bears would reduce the size of —0.107 —0.508 0.517
wild hog populations

Black bears in East Texas would 0.723 0.232 0.421
increase my quality of life

Black bears near my home would 0.738 0.225 0.423
increase my quality of life

Black bears have the right to exist 0.642 0.053 0.068
wherever they may occur

I would feel personally at risk if black 0.787 —0.413 ——0.070
bears exist in East Texas

I am afraid of black bears 0.765 —0.428 0.022

Black bears commonly harm humans 0.731 —0.41 1 -0.075

Wildlife experts know how to manage 0.129 0.885 0.153
black bears

Wildlife experts understand landowners’ 0.098 0.891 0.161

concerns about black bears

 

aAll initial variables were displayed on the survey in Likert scale format (5 = strongly
agree, 4 = agree, 3 = unsure, 2 = disagree, 1 = strongly disagree) and were coded for
analysis so that greater values indicated greater support for bears. Highlighted values

indicate variables that factored together.

bEliminated from further analysis.

114

we; u Om
”Nos .1. :82 6L 0 owcmm

Nm.m n Om
”2:: n :82 £7; .1. owcmm

$33 .1. oz ”fa—.2 u a;

sac. .N n 808 .5 256$

$6.3 n @365 8 ooodom
Sven u $3.3m 2 ooodvm
$5.3” u 9&de 8 ooodmm
an: n 25.8» as 33

flood u 8..ch $0.85 8 89680
9°: .1. 03:va owozou

$2. .l. 08on 983083.

ficwwm n owe—Bo oEom

cams u Bosom bush 8 Ecocmoo>
$va u 28.353 be 32.3 :3:
as; n 32.8 FEE

.72 am was; 3% u as:
aim bass“. 23:. n 222
seam u 02 3592 u 8>

mo; n Om Kod u :82
$0.3 n Efiéoc 283.
o\..¢.~_ u Edam .153
9°de 0 =39 :55
o\ov.m_ u :38 “0qu
$06 M 9:55

$98 u as :25

foo H .06 owns

:82 8228s 002.5
A: $52 no AS 883083 AC 5&0

:52 822% 58
3 ssz a .5 85858 .5 520

8v 02 a .3 8>

m n 22: 20 25.3%

v u @363 8 25.8%

m n oomdmm 2 25de

N n ooadmm 8 ooodmm

_ u 8°.on 55 33

n n 82ch 356835 no 89680
c 0 0:26me 03:00

m u 02on So» 95 Lo mb§8mm<
v n 03:00 oEom

m n .0023. 02E 5 Ecocmoo>

N u Ammo rwdv EuE>Sco no .0058 FE:

_ n 3 oumuwv .828. basin

3% 8E Ba 38 5253 8:28.26
:EE .8 So» 3235 Eoccoamom
§ 222 a .5 2552

8V 02 a .3 3>

Eoucoqgc .3 325E waS
n u Shaina: .ESM

c n 8.5 .ESM

m u E88 25.3 :39 :55

a n 3.8% 25.2 i 25.3 :32 awe;

m M £35
N u 3.83 8qu - 80.21% :25
_ n 8303 ooodmAv .06 owned

8238a 03862390333
5 0392223 Seesaw“: of Boo
@0523» 9
3:22 3:338 Swims—c:
5 San—BEE Eon—5&2 ofi moon
9:262ng @822
02:23» m Co .382: a so» 2<

wmoom E «.28. 883 2:85
20:33: 8ch So» 83 8:?

2333388
022. so» 35 5:850
38.28 .3 .32 3.232 05 a 8:3

mEon so» 203 an» 22.3 E
magnum do 22: so» 2<
mama >5 02:. so» 00
2:0 Emu» Ev Pa @3530:
So» 5 29.239: .39: So:

moz— onES :0»
on 362:an mo 09¢ :23 5

8338s
82$ aozfiooammbzmmmm
2.823 £83.58 55%:

Qmov 2882: :ozmficmwco

2.323 2:85

22% caesium
Gag aw...
$3 saw
233 sum

3% iv 398:2

Ammov 09$ bE=EEoU

 

43.32 ozatomoo

“568385 Co wEuoU

cosmosc xotsm

3 bzaE>

 

.238... “mam E cotflzaoa 33 x83 05 E 0822: an
25 .mbwwcmE 8:2; .383 183 2958 325:3 mo marge 5.2 saw a :88 can mowflcoobqv 3.2mm: o>catomon 35 com: 33mins 503326:— .N.v 033.

115

.m 2...... c. 322%... o... 5...? .o 3.38. 2.. £6.22... .88.. ..o 822. 83 292.? :28 0. 8.23.2. 3.22-2523 332...... ..o .cchw.mm<..
3:228. .82 ..o w:..a...mm.o>o .0. .5508 0. 3.3.25 0.53 3.28. 3.3.830.

 

$06.. n 02 ”$3.8 H mo>
o\om.m. u 530m

.o\co.om H 53.: 3....di u .83.
E... u Om

.3 u .meV 88.8 wvd u 5.22
Sew H Om ”3.2. CNN n :83.
0.3 u oz 2.... n 3.»

mm... H mm .28.. ”SM u :32

a... u am so... u :82 «I u “was.

.8 0z .o .5 3..
5 58m .5 AN. :85 .5 .83.
o-onow..m. ”awnogocx .80. u 85

8:83.. some .8 8. 02 .o 2. mo>
E22830. .3 .2223... .09....

8. oz 8 .. .v 8>
E2533. .3 323... .09....

2...... 3.23.2. 0.5.5
2. .262 .0 Am. 85.88% Am. :2.0

$2.3 2..
c. 32. 22.... a zoom .o>o so» 32.

82.8% 29:3 :0 88m

Enos Sons
£5.52... x.m 8 822.8. co 3mm
$.30 so» on can. .0 8.2. is... .53.
$8.83. 9.2 .5388. 2.. 800
$88... .mmm
:. .53. so» 022. v.32. >52: Bo:
$2.2...» o.
3.20. 8.22.2. 3.28-2.0?
:. 2222...... €22.38. 2.. 300
3:6...» o. 3.22

a m... 32. 8m

8%.. v E:.m..m

.33 o.w..........55.
A33 .235 8.9...
R3. 200.83..

3.3 8... E 2....

$3.. .352... 8.22-...03

3......00. .N... 2an

Table 4.3. Results of factor analysis (principal components analysis with varimax
rotation), used in data reduction, for grouping activities related to wildlife that

respondents may participate in (N = 972).a

 

(Percent variance explained)

Participation in Activities and factor loadings for data reduction

 

 

Utilitarian Passive- Work-related Otherb Otherc
appreciative

Activity (32.4%) (9.2%) (8.2%) (7.3%) (6.4%)
Hiking 0.175 0.330 0.120 0.575 —0.121
Running 0.009 0.018 0.168 0.740 0.276
Biking 0.097 0.022 —0.060 0.782 —0.055
Camping 0.699 0.152 0.128 0.090 —0.143
Boating 0.796 0.056 -—0.064 0.148 0.062
Canoeing 0.385 0.088 0.028 0.345 —0.503
All-terrain vehicle use 0.521 0.118 0.431 0.157 0.287
Read about wildlife 0.202 0.809 0.126 0.085 0.051
Watch wildlife-related 0.143 0.832 0.040 0.028 0.074

TV or movies
Observe wildlife 0.173 0.766 0.142 0.137 —0.002
Hunt big game 0.601 0.287 0.432 —0.100 0.348
Hunt small game 0.582 0.279 0.376 0.017 0.371
Fish 0.749 0.251 0.069 0.043 0.059
Work on a farm 0.115 0.229 0.674 0.087 0.006
Work for the timber 0.051 —0.007 0.782 0.033 —0. 143

industry
Work for the oil 0.208 0.087 —0.101 0.142 0.669

Industry

 

8Respondents were asked to indicate “often” (coded as 3), “sometimes”(2) or “never”

(1) based on individual participation in each activity. Highlighted values indicate

variables that factored together.

b‘cEliminated from further analysis.

117

macs x83 Son:

 

 

65.8 8.56:8 .mbgoufiw—
mod ad we: mém Wow «.0 03 @5353: 3.898 £225?
85.8 $3: x33 omega: 3:33:
cud o._ mé odm :m 5.2 :0 9 Bo: >55— mtomxo cit—2:5 60:59:32
Amwdv 35:5:
N3. 3: a? n2 3. 2 MS :92 28868 can as:
a _ . :
mm.m ww— w.mm odm n. a we m3 383 xoflnmo 23% Ed;
*3.on
Ame. S “mam E E5 mason x83
66 Z: of: odm N; _ ad 33 t fit 3 >=m:ofloq Eon 2:03 H
8o. C 58: >2: 565:3
mm.m Y: :0— o.C m6: 7: m3 3:8 8 Emu m 96: Emma xofim
$0.8 om: mo >595 >E ommosfi
ohm >42 mNm Ohm Wm— w.m to 2:95 2:0: >8 :3: $8: xoflm
80.8 om: mo >523: >8 338E
3: as flow ob: ”2 mm m3 2:03 9“on “mam :M £39 Jon—m
Amwdv 608:835 >530: a :0 3:39:
m>.m : g m. H m :0: 0: mg :me 2 mason x83 («0 8:085 .83 xoflm
E Q 5 c; 5 8:83 33% o§§>
wewmflw 02w:
>_w:obm oflwmma 83:3 089:. >_w:obm
am:
:82 830:8: go gov owfififlom Z

 

.3on 583238 E 5:268:58 Eon xomE a 2:38 8:35: $33 0: wow: 3:028:
8:: mo wEEEwEoS :8 E508: 9 Bfiwfiav %oS£8> 66:53: mo 25 H :38 “E: 838855 8:58 oZEtomoQ 6.: 2an

118

AL V255: :: >9 @3865 2: .8396: QB 38:: 03:3 2:39 3:39: 0333:: 208 8:065 mo=_:>
$sz :5 :03 >o:8m_m:oo Hob wfino: £033 :8 £5383: >o>.5m :53 :33: 2:38 8:35: 0383:: 22: m:::o€:_ m02:>
5:an 53> .mo:_:> 83:30: E::::% ::: :88 3:328 8 tom: 895 @2833: 826:: 86:: 88:32:: :3 mo:_:> 2:8 toxin

 

8:35:
Co. 3 33803 on 230% 36:. 3:805
mm: .2. a: 8.9. EN 4.: m a 8m a 88:88 88 8:3 2:. 8838:

3.88 .3. 28:

119

30:38.8: Eo::8:om :08: 00:88::me 88:0: ft ::: ._0>0_ 00:0:830 $3 08 :: 00:88::me 88:0: AL:

 

 

...... 83w .2: .N 8:8: "L: 88 8 .N 9.2 u L: :88 E. .N 83 u: 88 8m
2:: .88 .m as: u p: .88.: mg .m an... n L: 88 :2. .m :3: n k :88:
L... 59% 88.: u L L... 82m E: n L L... Sodw 3.8 u L 08288:
88.: 88.? u L :8: 89:. u L :8: :88. n L :88 88¢.
38 .8: .: 82 u ...: an: :8 .: :8: u: as: :8 ._ 58 n: 8985
15.: 38:: u L .88.: 88:. n L L... 53% .88.? u L 88 a 08:
33 :8: u L £2 8:: n L 8.: 80.: n L 83:8 888383
>833:
L... 53m 82 u L isodw :2: n L :88 88 n L 88888.88:
.2. 82m 8:: u L 8:: :8: u L L... 59% 88.: u L m8:38.N 8:86:
L... Sim m a .: 38:: u L: ::: ma .: 38:: n k 88.: at. .: mg: n: $888 8:85:80
38% waduL 2:: :::”..L 18:: 9:: "L 288:
2:: 03.: n L mm: :::.o u L m8: as: n L 8:88:
L... 33m :8 : :1 n L :3: 82:. u L :3: 82:1 n L 08
L... 59% :3 .: :89. u L: :8: :3 ._ :8: u L: L... 53m S ._ :88 u: 8m
88.: w a ._ a: u L: :2: :8 .: 8:: u L: Leo: E ._ :2: u L: 8:
.18.: 2 S n L :8: So: u L .. so: :3: u L ::v 8:82
SN: :8? n L Q _ .o :88. u L 83 :8? u L 8.: 58880
R .3: 03$:me amok K wt otmfifim amok. K .2: ofimflwum amok. 033.:me “COUGQQQUGH

 

08085 33:38: :::33 8:33;:

m:0w:::E :::33 8:334:L

.803 Mom—n— UHNBOH mvfisfifififl

 

03:88:, ::0::0Q0Q

 

.88: 8:33: ::0: 8:3 88.: 50380838 3 :08 8E:_::> :8 a8m>_::: 30:83: 0::_::>E :8 3388 6.: 28¢

120

 

 

$3 $2 $3: 52 MS: :8: as: mm; :8: :8: 8m
:3: $2- mad: as: 32. so? 33 32- m3? 8385
.. 53m 82 mom: .. so: 33 2:: .. 89% :3: 83 0:355:
.. so: 33: mm: .o. as: £2- :3? m8: :3. T as? 8:30 8:2
2m: 9%.: 3.0.: $2 :8: m8: m3: 8: :8: 0:235
8:: 53. m3? 3:: SE- «8.? .. : a: $3- :::. 8L: a 25:
3.2 :8: So: E: $3. $3: $2 :2: :8: :::,qu 8:220:53
£33m

18.0w mm? 83 .83: :3 8:: 18% NR: .83 02:62:30sz
3 3 an: :8: 82 m : S 2:: 22 $2 :8: :38: 5:955
:3: so? Sod: 5:: 9.2:: 80:. 8:: an? :86. $3805 8:32:30
$2 :2 :8: 3:: fine: EST :3 $2: :5: MU58::
3:: :3. :8? 3.2 :3? :85: $2 two: :::. 8:83:
.83 $2- NE? E: $31 23:1 as: $2- $27 0?.
... 82m $31 23. an: :3: :8: 185w .83- :::.ou 5:
3w: 5.: moo: 3:: 8:: so: :2 :2: m3: 8:
:53 £2 :8: : : no so: So: an: 8:: 9:: w :v 5:832
:8: 3:. T ES] .58.: :::.Nl as? 82 0%.? as? 0%: £5888
: L M: L: L n m L m 03%? 20:58::

 

0N6 3:538: ::09 03:3
0:: 3:80:05 :::33 8:3:.:<

m:0w:::3 0.3—:23 :::33 8:33<

@509 Mow—D “Yak/OH mDUSU—mtdx

 

0E::::> ::0::0Q0Q

 

03: 33:33:: ::0: 06:3 0:: w:_m:0:0:_ ::: .um:0w:::::

02:23 .238: 8:3 :::33 8:33: :0: L0:oE 58830: 8::ch :80— 33:5 33060:: mix—::: 03:33:22 6.: 03:.H

121

:56 .v. a as. a n k “Gad n ma. ..an ”mac n m:

53 w m ”S3. u m A83 u E «as 33 u ”x0

:56 w a $5 u k “ammo H mm «$3 $2 n NE

.322 85338 $3 2: E 088$:me 88:3 AL :< .3282 $56508 $36828“?

Afibcouv .06 28$

122

 

 

      

San
Jacinto

 

 

San 3
Augustine

 

 

 

 

'11: ' E"
Angelina l

 

 

 

 

 

 

Newton E

J
.'l

 

 

 

 

 

 

 

 

 

 

 

’ Orange 5

 

 

Figure 4.1. The l2—county region in southeastern Texas used to identify attitudes toward
black bears, wildlife managers, and increasing the black bear population. Survey
participants were selected from these counties.

123

CHAPTER 5

SOCIOECONOMIC FACTORS AFFECTING SUPPORT FOR BLACK BEAR

RECOVERY MANAGEMENT STRATEGIES

In collaboration with

Angela G. Mertig, J ianguo Liu, and Nathan Garner

124

Abstract

Many studies have assessed public reaction to various wildlife management
scenarios. However, few have specifically focused on public opinion toward a
reintroduction versus a non-assisted (i.e., “let nature take its course”) recovery of a
species. We surveyed 1,006 residents throughout southeast Texas in an effort to
determine variables that influence opinion of preferred management strategy for recovery
(or non-recovery) of black bears. The general desire to have black bears in the area had
the greatest influence on selection of preferred management strategy. Concern (worry)
about problems that bears may cause was also of great influence. Most importantly, there
was a difference in predictive variables for two strategies involving human assistance that
differed only in their wording, but not their meaning to wildlife management. Therefore,
managers must be very clear when explaining details of a management strategy in an

effort to minimize miscommunication and confusion between managers and residents.

125

5.1. Introduction

During the past decade, increasing attention has been given to public attitudes
toward wildlife. A number of studies have assessed public attitudes toward different
management scenarios regarding human-wildlife interactions, and have provided
managers with information about how to best minimize negative outcomes of conflicts
between humans and wildlife. Examples of recent research include hunting versus non-
hunting in urban areas for managing elk (Lee and Miller 2003), accountability for
damage by geese (Coluccy et al. 2001), managing for predators to enhance prey
recruitment (Messmer et a1. 1999), attitudes and behaviors toward feral cats and preferred
methods of control (Ash and Adams 2003), acceptability of management actions as a
result of different types of encounters by mountain lions (Manfredo et al. 1998), defining
policy for black bears based on attitudes and opinions of residents (Peyton et al. 2001),
and ascertaining variables that determine preferred predator management practices (Teel
et al. 2002). Studies of acceptance by humans who have experienced damage or injury
by a particular species are also common (e. g., black bears; Pelton and Scott 1976; Jonker
et al. 1998).

Such knowledge is extremely pertinent to improving wildlife management
practices. Many studies have addressed public opinion about particular species involved
in potential reintroduction strategies (Bath 1987; Lohr et al. 1996), but fewer data are
available regarding public attitudes comparing management alternatives in such
instances. Regardless, it is almost impossible to determine how residents of an area will
react to human-wildlife interactions after a species is present, especially if the species,

such as the black bear, is perceived to be of potential negative consequence (i.e.,

126

perceived danger to humans). It is therefore essential to collect as much information as
possible to determine particular variables that may contribute to public reaction, and
provide managers with baseline data for comparison with post-reintroduction public
attitudes.

The Louisiana black bear (Ursus americanus luteolus) was extirpated, as a result
of over-harvest and habitat destruction, from most of its historical range in Louisiana,
southern Mississippi, and southeast Texas by the early 1900s (BBCC 1997). At the
beginning of the 19905, only two small remnant populations remained in eastern
Louisiana. An extensive public outreach and information campaign took place in
Louisiana recently to build public support for bear recovery throughout the state. Two
recovery plans were created (Bowker and Jacobson 1995; BBCC 1997), and recovery has
been successful thus far (Van Why, 2003). To date, feasibility analyses have also been
completed in Mississippi (Shropshire 1996; Bowman 1999).

In eastern Texas, there has been an increase in the number of confirmed black
bear sightings since the early 19908. At this time, there is no known breeding population
in eastern Texas, and sightings are most likely transient bears from Arkansas, Oklahoma,
and/or Louisiana populations who wander into eastern Texas. The increased number of
sightings in Texas has prompted the creation of a Texas black bear conservation and
management plan. Objectives of the plan for the next 10 years include public
coordination, communication, outreach/information dissemination, habitat management,
and research (TPWD 2005). The ultimate goal, through partnerships between natural
resources organizations and the general public, is to restore habitat for the purpose of

reestablishing black bear as a viable ecosystem component in East Texas (TPWD 2005).

127

The objective of this research was to ascertain public attitudes toward
management for recovery of the Louisiana black bear in southeast Texas by comparing
support for alternative management strategies, and to determine the socioeconomic
variables of greatest import for understanding these attitudes. We assess support for three
general management options for black bears (described in methods section): (1) human-

assisted reintroduction, (2) natural (non-assisted) recovery, and (3) no bear recovery.

5.2. Methods
5. 2. 1. Study area and survey design
Please refer to Sections 4.2.1 (including Figure 4.1) and 4.2.2 for descriptions of

the study area and survey design, respectively.

5. 2. 2. Dependent variables

We selected four dependent variables to represent three different potential
management strategies. Two of the scenarios, human-assisted and restocking, were
conceptually equivalent to us, but we sought to determine whether vernacular used would
affect responses. For the first three variables (discussed below), we asked participants
whether they support, do not support, or are unsure about the specified strategy. These
three questions were coded such that higher scores indicated more support for a particular
management scenario (support = 3, unsure = 2, no support = 1). For the Natural and
Assist variables, we asked respondents to also volunteer reasons for their selection (i.e.,
support, not support, unsure). We then categorized similar responses to quantify

rationales for each selection.

128

Natural. —Respondents were asked whether they supported the following
strategy: “Do you think black bear populations in East Texas should increase naturally
(i.e., without assistance from a natural resource agency)?”

Assist. ——Respondents were asked whether they supported the following strategy:
“Do you think that natural resource agencies should assist in increasing the black bear
population size in East Texas?” Specific means of assistance were not provided.

Restock. —Respondents were asked whether they supported the following
strategy: “Would you support the restocking of black bears into suitable habitats in East
Texas by natural resource agencies?” The Assist and Restock variables both sought
essentially the same information, but used different terminology. Restock is a
management term closely associated with game harvest (Bolen and Robinson 1999),
whereas Assist is a more general and widely used term related to species reintroductions.
We compared the results for both Assist and Restock to determine whether vernacular
differences between terms would result in differences in support.

N0 bear. —Respondents were asked whether they agreed or disagreed with the
statement “black bears should not exist in southeast Texas.” This management strategy
would involve the complete exclusion of bears from the area, and was asked using a
Likert—scale format. Responses were coded so that higher values indicated more support
for excluding bears from the area (1 = strongly disagree, 2 = disagree, 3 = unsure, 4 =

agree, 5 = strongly agree).

129

5. 2. 3. Independent variables

We selected 21 independent variables for our analysis. Seventeen of these
variables were the same as those described in Chapter 4 (Table 4.2), and four additional
variables were used (Table 5.1). Based on findings and reasoning from similar studies,
we hypothesized that male, younger, more-educated, higher-income, and more
knowledgeable (about bears) respondents would be more likely to support a human-
assisted means of recovery (Assist and Restock). We expected that members of more
heavily populated community types, as well as those who live in the south and urban
strata, people who have seen bears in the wild, respondents who want bears in East
Texas, newer residents to the area, respondents who are interested in wildlife, members
of a wildlife-related organization, residents who participate in activities related to wildlife
more frequently, and residents who have lived in areas where black bears are present
were also more likely to support human-assisted means of recovery. Finally, we
hypothesized that residents with children, pet owners, large-tract landowners, respondents
who would worry about the problems that bears may cause, and individuals who tend

livestock would be more supportive of natural or no bear strategies.

5. 2. 4. Non-response follow-up

Please refer to section 4.2.4 for discussion of the non-response follow-up survey.

5. 2. 5. Statistical analysis

Please see section 4.2.5 for descriptions of statistical analyses. Recognizing that

our dependent variables are ordinal and most have only 3 categories, which technically

130

necessitate the use of ordinal logistic regression (Babbie 1990; Sokal and Rohlf 1995),
we ran the same multivariate analyses using ordinal logical regression. We found no
differences in terms of which parameters were statistically significant and in what
direction. Similarly, parameter values were strikingly close in magnitude to those arrived
at using regular multiple regression. Because readers are likely more familiar with
multiple regression, we chose to present the results as such. All alpha values were
defined at the 95% confidence level. After accounting for multiple comparisons in
bivariate analyses (21 tests per dependent variable) with a Bonferroni correction, the P

values were considered significant at the level of 0.002).

5.3. Results

Detailed results of univariate analyses of independent and dependent variables for
all respondents are displayed in Tables 4.4, and 5.1, and 5.2, respectively (see also
Morzillo et al. 2005). See Chapter 4 for a review of the 17 independent variables used in
both chapters. For the four independent variables unique to this chapter, 92% of
respondents indicated that they were interested in wildlife, and 65% of respondents
indicated that they wanted black bears in southeast Texas. However, 34% of respondents
who wanted bears worried about problems that bears may cause. Overall, half of
respondents indicated that they would worry about problems that bears may cause
regardless of whether they want or do not want bears in the area. Thirty-two percent of
respondents have lived in a location within the present geographical range of black bears.

Pearson correlations revealed moderate statistically significant correlations

between dependent variables (Table 5.2). A positive correlation existed between Assist

131

and Restock, which to us described essentially the same strategy, but were interpreted
differently by respondents. Correlations between all other pairs of management strategies
were negative. Support, non-support for, and unsure opinions about both Natural and
Assist were indicated by approximately one third of respondents each (Table 5.2).
Restocking the black bear population by natural resource agencies (Restock) was
supported by 50% of respondents, whereas the remaining half was almost evenly divided
between non-support and unsure. Only six percent of respondents held attitudes
suggesting that black bears should not exist in East Texas, whereas >70% disagreed or
strongly disagreed with the same strategy.

We completed bivariate analyses of each independent with each dependent
variable (Table 5.3). Natural did not vary by community type, whether or not the
respondent owns pets, age, education, how often the respondent participates in work
activities related to wildlife, duration of residence in southeast Texas, whether the
respondent tends livestock, the number of acres owned by the respondent, stratum, or
whether the respondent has seen a bear in the wild. Whether a respondent was interested
in wildlife, would enjoy having black bears in the area, and whether or not the respondent
had lived in a location where bears are present were also insignificant for Natural.
Respondents in households with fewer members <18 years of age, females, and those
with lower incomes, were more likely to support a natural increase in the black bear
population. Also more likely to support a natural increase in the bear population were
members of wildlife-related organizations, respondents who participate less frequently in
utilitarian and passive-appreciative activities related to wildlife, those with lower

knowledge scores, and those who worry about the problems that bears may cause.

132

Assist did not vary by community type, number of household members <18 years
of age, whether or not the respondent owns pets, education, income, frequency of
participation in work activities related to wildlife, number of years the respondent has
lived in southeast Texas, whether the respondent tends livestock, number of acres owned,
stratum, whether the respondent has seen a bear in the wild, or whether the respondent
has lived in a location where bears are present. However, females, younger respondents.
members of wildlife-related organizations, more frequent participants in utilitarian and
passive-appreciative activities related to wildlife, those with higher knowledge scores,
those interested in wildlife, those who do not worry about the problems that bears may
cause, and those who would enjoy having bears in the area were more likely to support a
human-assisted bear recovery program.

Restock did not vary by community type, whether the respondent owns pets,
education, whether the respondent participates in work activities related to wildlife, the
length of time that a respondent has lived in southeast Texas, whether the respondent
tends livestock, number of acres owned, stratum, whether a respondent has seen a bear in
the wild, or whether the respondent has lived in a location where bears are present.
Respondents in households with a greater number <18, females, younger respondents,
respondents with higher incomes, members of wildlife organizations, respondents who
participate more often in utilitarian and passive-appreciative activities related to wildlife,
respondents with greater knowledge about bears, those interested in wildlife, those who
do not worry about the problems that bears may cause, and those who would enjoy
having bears within the area were more likely to support restocking of black bears into

suitable habitat by natural resource agencies.

133

N0 bear did not differ significantly by community type, number of household
members <18 years of age, whether or not the respondent had pets, sex, frequency of
participation in work activities that are related to wildlife, whether respondents tend
livestock, number of acres owned, stratum, and whether the respondent has lived in a
location where bears are present. Older respondents, respondents with lower levels of
education, respondents with lower incomes, respondents who are not members of a
wildlife-related organization, those who less frequently take part in utilitarian and
passive-appreciative activities, those with a longer duration of residence in southeast
Texas, and respondents with less knowledge about bears, were more likely to support
management calling for no bears in Texas. Respondents who have not seen bears in the
wild, do not have an interest in wildlife, worry about problems that bears may cause, and
would not enjoy having bears in the area were also more likely to support the no bear
management strategy.

Multivariate analyses yielded a regression model for each dependent variable
(Table 5.4). Households containing fewer individuals <18 years of age, respondents who
are less likely to participate in utilitarian activities related to wildlife, those who own a
greater number of acres, and those who are more likely to worry about the problems that
bears may cause, are more likely to support a natural increase in bear populations. Men,
respondents who participate more often in passive-appreciative activities related to
wildlife, those who do not worry about the problems that bears may cause, and those who
would enjoy having bears in the area were more likely to support a human-assisted bear
recovery. Respondents who participate more often in utilitarian and passive-appreciative

activities related to wildlife, were less worried about the problems that bears may cause,

134

and would enjoy having black bears in southeast Texas were more likely to support
restocking of black bears into southeast Texas. Respondents who have lower incomes,
have lived in the area for a relatively longer period of time, have less knowledge about
bears, worry more about the problems that bears may cause, and would not enjoy having
bears in the area were more likely to indicate that black bears should not exist in East

Texas.

5.4. Discussion

Our objective was to determine whether support for particular black bear
management strategies is influenced by particular demographic and socioeconomic
variables. To our knowledge, this is the first study to compare how respondents’
characteristics influence selection of preferred management strategy for implementation
of a species recovery. The regression models performed well, predicting 17-51% the
variance. Independent variables predicted approximately 17% of the variance of Natural.
Although Restock and Assist sought the same information from a management standpoint,
there were some differences in predictor variables. By asking both, however, we have
learned that respondents interpret these terms (i.e., potential management strategies)
differently. Respondents may have interpreted Restock as placing more bears into
specific locations within the area, and Assist as a strategy that may involve more-rigid
environmental regulations. For future information dissemination, wildlife managers must
explain potential bear management implementation plans in detail and avoid any

confusion or potential misunderstandings with local residents.

135

The general desire (or not) to have bears in the area, and whether a respondent
worried about the problems that bears may cause, were the independent variables that
most consistently influenced dependent variables, when controlling for all other
independent variables. In other words, a person’s general attitude toward wanting bears
in the area (whether they would see them or not) had the most influence on whether a
respondent supported an active (human-assisted; i.e., Assist or Restock) or No bear
management strategy. Similarly, respondents who worried about problems that bears
may cause were more likely to support the passive Natural strategy, or prohibiting bears
from establishing themselves in southeastern Texas. Future studies must delve even
deeper into the reasons why people have positive or negative attitudes toward bears in the
first place, as the socioeconomic variables used here did not make the effect of these
attitudes disappear.

In our study, the desire to let nature take its course, concerns about funding and/or
the opinion that bear recovery was a waste of money, and the desire not to have bears in
the area were most often cited by respondents as reasons why they opposed Assist.
Conversely, the most commonly cited reasons for supporting Assist were that natural
recovery of the bear population would take too long and respondents wanted to see or
have more bears in the area (n = 38 and n = 36, respectively). If local residents want
bears, they may consider the presence of bears to be a greater benefit than concerns for
costs related to negative impacts (e. g. nuisance activity).

When respondents were asked to identify reasons why they might oppose Natural,
it became clear that much of the opposition to this strategy stemmed not from opposition

to bears, but rather a desire to see bears even more quickly than natural processes would

136

allow. Many respondents indicated that they did not support Natural because active
recovery of bear populations was important for a better chance of survival (n = 98) and
for the enjoyment of future generations (n = 3). Within the general comments section of
the survey, several respondents indicated that they would like their children to experience
the possibility of seeing a bear. This rationale is further supported in our bivariate
analysis, in which respondents belonging to households with a greater number of
members <18 years of age were more likely to prefer management strategies that
involved human assistance. Just the opposite, respondents from households with fewer
members <18 year of age were more likely to support the N0 bear or Natural
management strategies. However, the number of household members <18 years of age
remained significant in multivariate analysis only for Natural, suggesting that the effects
of this variable occur through other variables included in the model. Support for a
reintroduction because the area is within the historical range of black bears, and because
of the significance of bears to the area’s culture was mentioned (n = 5), but was not as
prevalent as in other studies (Enck and Brown 2002).

Worry was significant for all four dependent variables. People who worried less
were more supportive of human assisted recovery, whereas people who worried more
were more supportive of Natural or No bear strategies, even when controlling for other
variables. However, from the reasons provided by respondents in our survey, concern
about problems that bears may cause was only the fourth most common reason for not
supporting any of the management scenarios that would result in bear presence (n = 17).
More often indicated were let nature take its course (n = 86), concern about monetary

cost of a bear recovery (n = 31), and simply no desire to have bears (n = 29). Lower

137

income respondents, as well as respondents who have lived in the area for a longer
amount of time, may be concerned about paying for bear damage that might occur on
their property. In fact, lower income individuals were more likely to prefer the Natural
and N0 bear management strategies. This may include elderly residents who live on
fixed incomes and are possibly also concerned about their own safety, as our results
indicated a significant correlation between age and income (r = —0.202, P $0.001), as
well as between worry and age (r = 0.094, P = 0.008). We also expected that respondents
with pets would be worried about pet-bear conflicts. However, whether the respondent
owned pets was only significantly positively related to Restock, and only before
application of the Bonferroni adjustment, which may suggest that respondents are aware
that bears are unlikely to harm pets. In fact, concern for pets was mentioned as a reason
for non-support for only Restock, and only by one respondent.

Concerns about both direct and indirect human-wildlife conflict are not
uncommon. Perceived negative effects on livestock were positively related to non-
support for a wolf reintroduction in New York (Buck and Brown 2002). In Arizona, elk-
vehicle collisions and feeding by humans were mentioned as potential problems with
urban elk populations (Lee and Miller 2003). But human-wildlife conflict does not
always result in human intolerance for a species. In Massachusetts, bear damage to crops
and livestock was considered only a slightly to moderately intolerable inconvenience
(J onker et al. 1998). Even individuals who experienced bear-related damage in one study
(<5% of those surveyed) considered the presence of bears as an overall positive impact
on the area (Jonker et al. 1998). Wildlife managers in southeast Texas must seek to

determine whether local residents’ attitudes are likely to change if bear-related damage is

138

(or is not) experienced (i.e., Social Carrying Capacity; Peyton et al. 2001). Proactive
solutions may then be established to resolve situations involving residents who
experience bear-related damage (e. g., reimbursement program). For example, in
Massachusetts, most bear problems involving livestock took place during the birthing
season (J onker et al. 2003). This is concurrent with the time of year when bears, which
are opportunistic foragers, search for food for cubs and require extra energy for nursing
and recovery after hibernation (Lariviére 2001). Electric fencing was determined to be
the most successful proactive means for deterring bears from livestock (J onker et al.
2003). Therefore, regardless of management course of action, we suggest that wildlife
managers incorporate suggested strategies for proactive bear deterrence (e. g., electric
fencing) into public outreach efforts (Colluccy et al. 2001; Jonker et al. 2003).

Outreach by wildlife managers should also seek to increase the public’s general
knowledge about bears, which may assist residents in making decisions that lead to
minimal bear-human conflict. Respondents with less knowledge about bears were more
likely to support N0 bear in both bivariate and multivariate analysis. In our survey, lack
of knowledge about bears (or not enough knowledge about bears to make a decision) was
the most cited reason across all management strategies by respondents who were unsure
about their preferred management strategy (n = 156).

Higher knowledge scores were significantly positively correlated with more
frequent participation in passive-appreciative activities related to wildlife (r = 0.422, P
30,001). Like respondents with higher knowledge scores, respondents who participate
more frequently in passive-appreciative activities were more supportive of both Assist

and Restock. Thus, the mere presence of bear may hold important existence value

139

(Boardman et al. 1996) or emotional fulfillment to individuals who more frequently
pursue passive-appreciative wildlife-related activities. This may partly explain why more
frequent participants in utilitarian activities related to wildlife were more likely (in
multivariate analysis) to support Restock, but not Assist, as restock is often related to
harvest (Bolen and Robinson 1999). However, enjoyment of the outdoors and observing
wildlife are also common reasons for hunting (Decker and Connelly 1989). Regardless
of whether passive-appreciative participants hunt or not, they may participate in other
activities defined as utilitarian (i.e., camping, boating; r = 0.467, P 50.001 for
comparison between passive-appreciative and utilitarian participation). Respondents who
participate less frequently in utilitarian activities were more likely to support Natural,
suggesting that supporters of this management scenario may be more preservation-
oriented, preferring natural recovery to any interference from humans, positive or
negative (Pacelle 1998).

Several variables that we expected to be significant in multivariate analyses were
not significant. First, greater knowledge about a species has been associated with both
more positive (Kellert 1985b) and more negative (Mertig 2004) attitudes about the
species. Our results support the former via a positive bivariate relationship between
knowledge and Assist and Restock, and a significantly negative relationship with Natural
and No bear. But, the negative relationship only held true between knowledge and N0
bear in multivariate analysis, possibly for the reason that knowledge about bears is
correlated with the desire to have bears in the area (r = 0.289, P < 0.001), as well as

participation in passive-appreciative (r = 0.422, P < 0.001) and utilitarian (r = 0.267, P <

140

0.001) activities related to wildlife. Unfortunately for pro-bear advocates, future efforts
to increase knowledge do not guarantee changes in attitudes (Enck and Brown 2002).

Education and gender are two variables that are often associated with differences
in attitudes about reintroductions and other wildlife management situations. Similar to
results from other studies (Kellert 1985b; Lohr et al. 1996; Bowman 1999; Bowman et al.
2004), we expected that respondents with lower levels of education would be more
supportive of No bear, and less supportive of the other strategies. Education was
negatively significant for only No bear, and only in bivariate analysis. In bivariate
analysis, females were more likely to support Natural, Assist, and Restock. However,
results from multivariate analysis suggested that males were more likely to support
Assist. We expected that males would be more supportive of all three recovery strategies
that would lead to an increase in the number of bears, as males were more interested in
hunting bears than females (Morzillo, unpublished data). Results have been mixed in
other surveys involving carnivore reintroductions; men have been more supportive than
women (Kellert 1985b), women have been more supportive of men (Lohr et al. 1996), or
no significant difference between the sexes has existed (Shropshire 1999). Regardless,
females were underrepresented in our sample, and their opinions should be persistently
sought

Assuming that membership in wildlife-related organizations is related to interest
in wildlife, we expected that members of wildlife-related organizations would be more
supportive of management strategies involving human assistance. Membership was
significantly related to support for Natural, Assist, and Restock in bivariate analysis only.

Because members of wildlife organizations were more likely to want bears in the area (X2

141

= 4.985, 1 df = l, P = 0.026), it is possible that the bivariate significance became masked
in multivariate analyses by the strong influence of desire to have bears.

We expected respondents who tend livestock to be less supportive of pro-bear
management scenarios (Bath 1987). However, tending livestock was significantly
positively related to Natural, Assist and Restock in bivariate analyses before Bonferroni
adjustment, but not significant in multivariate analysis. With no recent reporting of bear
harassment of livestock in East Texas, it is possible that livestock owners do not consider
bears to be a threat to livestock. Belief that bears are not dangerous (Kellert 1994) may
also support bivariate results that suggested more support for Restock and less support for
N0 bear among respondents who have seen a bear in the wild. Again, these relationships
may have diminished in multivariate analysis because seeing a bear is significantly, but
very weakly, related to a desire to have bears in the area (r = 0.093, P = 0.007). We
expected that respondents who have lived in areas where black bears are present would
be more likely to support pro-bear management strategies, but this variable was not
significant in any instance. Even if a respondent has lived in an area within the black
bear’s present range, he/she may not have been aware of the presence of bears,
particularly within urban areas, unless the local media drew attention to the presence of
bears for a particular reason (e. g., sighting in an urban area). Having more detailed
information about respondents’ past residences (e.g., city, rural area) may allow for better
evaluation of this relationship.

Neither community type nor stratum was significant in any instance. As in past
studies (Lohr et al. 1996; Bowman 1999), we expected that residents of more-urban areas

would be most supportive of human-assisted recovery strategies and rural residents of

142

anti-bear strategies. In our study area, the rural population is growing, mainly as a result
of emigration from nearby cities (e.g., Beaumont, Houston). This may explain positive
relationships between Time in area and both Natural and No bear, as well as negative
relationships between Time in area and Restock, but this trend does not hold true for
Assist.

Finally, we expected respondents who own more acres to be less supportive of a
bear recovery. Kellert (1991) and Enck and Brown (2002) both suggested that rural
landowners may be concerned with federal or state restrictions limiting land use as a
result of endangered species presence. This may be a reason for the number of acres
owned to be significantly positively related to Natural in multivariate analysis, but this
relationship was not present in bivariate analysis. In fact, only one respondent to our
survey noted government intrusion as a reason for not supporting bear recovery.

Assessment of additional demographic variables must be completed during future
data collection. For example, >25% of residents within the study area are of minority
race/ethnicity (US Census Bureau data). Approximately 96% of our survey respondents
were white/Caucasian. This results in a severe underrepresentation of minority
racial/ethnic populations in regard to management decisions that will affect these
demographic groups. In addition, because almost three-quarters of our respondents were
male, more attitudes data must be collected from female members of the population.
Better accounting for underrepresented groups such as those mentioned above will result

in a more accurate assessment of attitudes toward black bears across the study area.

143

5.5. Management implications

Although significant bivariate relationships existed between many of the variables
discussed, controlling for all variables caused many of the significant bivariate
relationships to disappear in multivariate analysis. In addition, further analyses will be
needed to assess present underrepresentation of other demographic characteristics (e. g.,
underrepresented minority racial/ethnic groups and females). Regardless, our results
provide important implications to wildlife management and more particularly to bear
management. In summary, most of our respondents indicated an interest in wildlife.
However, this interest does not directly transfer to a desire to have black bears in East
Texas. Desire to have bears in the area and worry about the problems that bears may
cause were the variables that best predicted respondents’ attitudes toward each
management strategy. It will be critical for managers to monitor relationships between
these variables over time, particularly as initiatives for a bear recovery progress. For
example, attitudes toward bears could change if the media broadcast bear-human
conflicts, especially if conflict involves a resident who is not supportive of having bears
in the area. At present, because of lack of bear presence, and the fact that extensive
outreach between resource agencies and residents is just getting underway, residents in
the area may not perceive bears as potentially influential in their lives (Enck and Brown
2002). For residents who worry about the problems that bears may cause, wildlife
managers need to further determine specific reasons for these concerns. In all instances,
it is essential that managers publicly explain in detail all management-related
terminology because any miscommunication or confusion of expectations between

managers and residents could lead to public distrust. As noted, there was a 20%

144

difference in support for Assist versus Restock which may lead to confusion if the terms
“reintroduction” and “restocking” are used interchangeably. As another example, the
term “assistance by natural resource managers” as related to a bear recovery could mean,
among other things, either managing habitat only or actively releasing bears brought in
from other locations. Such detailed explanation of strategies will allow for creation of
management practices that consider already existing concerns about bears in attempting
to minimize potential conflict between humans and bears in the context of specific land

[1868.

145

.3528.— Eh: mo 3:95.303 8m 5:88 8 @8563 803 3:58 ozatomofl a
.m 98 v 88930 flop 5 wow: 832E? E85332 C 05 How men 2an 8? com a

 

flames x83
mo emcee 58th 05 5:23 93: we:
Eugenia 2: 5523, 05::on

 

80.3 u 02 ”91.?” n 8% SV 02 no A: 3.» cc 33:5» 32 25382 $50 H23 5 o>3

.xsohm H 02 Sign H mo> SV 02 Ho A: 8% ©8on “mam E flees x83 :83 so» on 28>?
@833 $8 menus “m5

famdm H oz 35de H 3% SV 02 do A: 3% $5305 2: Sons who», so» 2:03 903

{cow n 02 maxeodo H mo> 8V 02 8 A: 8> wot—23> E 3388.: so» 2< 3.285

owe—38 oZEComoD EeEoSmeoE mo wEvoU cosmosc xofism A5 oEfiS>

 

.8on “mam E cote—smog Son x85 05 E 3385
cm was .mbwmcfic 056:3 .339 0333 @838 395:8 .6 £92.28 .28 smzsmoe guatomov can Em: 832%? Eot:og%£ 4% 03m...

146

Table 5.2. Descriptive results of explanatory dependent variables (weighted to account
for oversampling of rural residents) used to designate particular black bear recovery
strategies in East Texas.

 

Variablea (Management strategy; n) Descriptive results (% respondents)

 

Natural (995) Yes = 38.4; N0 = 30.2; Unsure = 31.4
Assist (986) Yes = 32.1; N0 = 36.2; Unsure = 31.7
Restock (991) Yes = 49.7; N0 = 22.5; Unsure = 27.7
No bear (981) Strongly agree = 2.1; Agree = 4.1; Unsure = 20.8;

Disagree = 41.1; Strongly disagree = 31.8

 

aResults from Pearson correlation (r) between the dependent variables are as follows (all
significant at P < 0.001): Natural—Assist = —0.553; Natural—Restock = —0.345; Natural—
No bear = —0. 120; Assist—Restock = 0.654; Assist—No bear = —0.434; Restock—No bear =
—0.524.

147

.8083:qu Bates—om 8cm cosmoficwfi 885v 0.19 can ._o>o_ 353:8 $3 2: an ougchwfi 88:28 AL.

 

 

 

E .o N8 ._ mNoN u a 32 N N35 u “N 5 .o N .4N_ .v n «x 32 N N2: n ma 2:
.. 5% N 83% n a». .. 83w N 335 n ..x 3.89% N 2 N: n «N :86 N 33 u N s53
.. Sim .N» .m 3.2: u a .. Seem N 822 n ...k .. 33w N E3: n a». .. 53w N 89% u 24 Ea?
.. 83m 3s ._ 33% u a ....._oo.ow N $4.2 n ”N :_8.o.v. N SEN n wk .32. N eNN.: n X mass
.183 so .N News N a .83 v 53. u ..x 83 v 2.3 u ..N 83 v 9.3 u E as. 8m

NM: 02 .N :2 u a 32 e 82 u k £2 0 83 n a». £2 e 83 n as. seen
.. 53m 221 n t ...... 33w NNNd u t .. 33m NNNe n t .. 53w 3;] n t sweeaoé

8N5 355 n t 83 God u t 22 83: n t $3 83 u t e256 as...

82 Sm ._ m: .o n a .83 N E: u a». ENS. N 5: u N .83 N 33 u ”N 8283..
......Sqow 3; .o u t 3:; mg? n t :55 £3. u t ..de .ng u t 8:. 5 25h

23 33. u t mN_ .o :3 u t N86 N86. u t 2 2 N8? u I assume “sausage;

m>30m
:5de SN? u t .18? «one u t .... 53m Nde u t .. 33w 3.31 u t 93823st
.. :53. Nm _ .o: u t .. 53w 3N5 n t .. 83w 2N5 n t .... 83w 8N5. u t messes 522:5
:8ch Ni ._ Ne: u a .. 59% N WSON u ”N .. 59% N SEN n N ......_8.ow N 2 _ .ON n a». tongue =2§§w5
.. 53w mm _ .o1 u t .. 52m :3 u t .83. :85 n t .. 83w NN_ .o. u t 0585
.... 83m 23.? u t mN_ .o 956 u t $2 83: n t :2 Noel u t Sesame
1.59% 83 n t ......Sqow Saint ......So.ow Edit 33 N85 ”L sm<
.83 a: ._ ONO.» u a 3.59% N 8va u a». .189 N 89:. n we ....._oo.ow N SNNN u a». saw
N35 a: ._ NSN u a Need N «8.2 n ..N 2N6 N 33. u «N 3.2 N 2 N._ n X as.
zoos 831 n t .. 53m 83 u t .33 N86 n t ......So.ow 2 3: u t N V 53:52
Nod 83 u t 3% Noel n t 3.3 Sod. u L we: mood. u t 25 @5588
Q ht Umwmmumam “mug. Q “U 03$“me amok K bu Uflmmwmwm amp—r Ax MU omummugm “wot—r 2333‘, uCOfin—DQUUG—

.503 OZ JQOHmoM Hmmmm< Ehamz

 

032:; ~528qu

 

.mbBebm EoEowmcmfi Eon x83 8m toga wEmmommm he now: 3325» no.“ .8335 28:23; 22.95 L8 3:53. .m.m 2an

148

53 w m .SNNN n k .233 u Nm ..e$ E 3 u E...
:3 w m 38.: n k .323 u "m :3; $3 u 3
53 w m “SN.” u k ASN.3 u "x ..as 33 u ”3
33 w m .23 u k #2 3 n ”m ..as 23 n F

._o>o_ 853:8 $3 05 as ougocEwfi 88:23 AL =< 68.5%: 356508 “336.829?

 

 

 

$3 NN3: No.3. 83 $3. :3. £3 83 83 N33 :3 83 2,:
18% ENT 331 183w ”N3 83 18.3 3.3 :3 N33 :3 83 :33
183w 33 $3 toqow $31 3N3] 18% So? 22.? 183w :3 83 :53
N33 33 NN3 $3 :2 83 $3 $27 83.? 3N3 82.. 3.? 3225
83 :3 :3 23 83 :3 :3 NN3 No.3 23 83. T 383. as 8m
83 M: 3 83 e 3 New: 83 83 NN3 383 e: 3 83. T SET 832%
$3 $3 .83 23 3.3- $3. $3 :3 «N3 .383 SNN 83 8533.2
183w 331 E3: :3 $2: 83.? :3 $2. $31 23 83 3N3 3238:
$3 83- E3: 23 :3 s3 83 £3 33 S3 331 33. 38:2,:
:3 NRN 83 33 $3; 23' $3 $3 3N3 83 :27 to? 8352.5
$3 83 83 N33 8: .o :3 33 NN3. 23. NN3 2 3: N53. beam“ 352:5?
£365

NN3 $3.? $3. .33 33 N33 :53 SNN $3 23 3N3: :3: 35850.3sz
NN_ 3 $2 $3 ...NN3 83 83 33 N2: :3 .33 231 £3. £23“ 5555
$3 ”on? :3: 23 2: $3 :3 :2 $3 :3 83- :3: 2352: =2§§w5
:3 $3: 23.? 3N3 :N._ 33 $3 3.3 $3 R3 :2. 83. 2:85
83 x3 33 $3 83.? 33: N83 $2: 38.? 83 $31 83. 8:82:
$3 23 $3 83 :3 83 :3 22- 3831 23 $2- $31 ow<
s3 :3 £3 £3 SET :31 ...23 EN. :31 E3 on: N83 5m
:3 83 N83 $3 83 S3 23 321 N831 93 $3 $3 son:
33 33.7 331 N83 :3: N83 83 N2: .83 18.3 32.? :3- 213532
33 No.3 53 33 R31 23. $3 N83 N33 :3 3N3. 33.? 25552.58

m N n Q: ~ n m 3 n m : n 03%? 202302:

oumon OZ uxooumom uummmm< 93.532

 

038:; Eovcoaoo

 

32335 EquwmcmE 53 x83 :8 toaasm .3533 .8 L085 coimewfl 33:3 :32 $590 wESEuE mfibacm 223332 in 28;

149

CHAPTER 6

SPATIAL ANALYSIS OF ATTITUDES TOWARD PROPOSED MANAGEMENT

STRATEGIES FOR A WILDLIFE REINTRODUCTION

In collaboration with

Angela G. Mertig, Jianguo Liu. and Nathan Garner

150

Abstract

Wildlife managers regularly incorporate human attitudes into decisions involving
wildlife conservation. Knowing the spatial distribution of particular attitudes may further
assist managers in determining distribution of support of or threats against wildlife
species. Using results from a mail survey and SaTScan 4.0, we assessed the spatial
distribution (clustering) of attitudes toward several management strategies for the
recovery of black bear in and around Big Thicket National Preserve, Texas. Statistically
significant clustering occurred for two attitudes: 1) non-support for a natural (non-human
assisted) increase in the bear population near the Angelina National Forest and 2) strong
disagreement toward total exclusion of bears from southeastern Texas within the
relatively urban Orange County. In addition, respondents closer to the preserve, a
potential black bear release site, were more likely to support exclusion of bears.
Analyses such as this can greatly assist managers in planning public outreach and

evaluation of social feasibility for wildlife reintroductions.

151

6.1. Introduction

As the human population of previously rural areas increases, there are an
increasing number of human interactions with wildlife. This escalation in human-wildlife
interactions has led to a new paradigm in wildlife management; rather than focusing
primarily on the ecological needs of species, managers increasingly appreciate the need
to incorporate human dimensions into wildlife management decision—making (Riley et al.
2002). As a key step in understanding the human component of management, many
studies have assessed socioeconomic and demographic variables that affect stakeholder
attitudes toward a specific species (Bath and Buchanan 1989; Kellert 1991; Lohr et al.
1996).

Attempts to elucidate attitudes about a species are exceptionally difficult for
reintroductions. Although researchers are able to identify potential habitat for a species,
it is nearly impossible to determine a priori whether (1) a species will use habitat that is
perceived as suitable, and (2) whether and how human-wildlife interactions will occur.
Insights about potential habitat may be gained from other ecological research involving
the target species. However, projecting results of human-wildlife interactions is much
more complex, and thus can only be attempted by gathering attitudinal and (anticipated)
behavioral information directly from people. Assessing human opinions of a species that
is a target for reintroduction, as well as how attitudes are distributed across space, may
allow researchers to identify locations where support or lack of support exists for a
species reintroduction. Spatial knowledge of attitudes may permit managers to identify
locations where humans’ attitudes are particularly favorable for a reintroduction (e.g.,

supporters may adapt land use practices for habitat management), as well as potentially

152

hostile locations where humans may pose a threat to the success of a reintroduction (e.g.,
harassment of the species).

Spatial analyses of data have been used for research in many disciplines including
public health (Knox, 2005), geography (Bailey and Gatrell 1995), human ecology and
agriculture (Naughton-Treves 1997; Sitati et al. 2003), environmental valuation (Brown
2005), and wildlife ecology (Mather et al. 1996; Kie et al. 2002). In the case of public
health, for instance, researchers seek to link morbidity rates to environmental factors. For
example, Knox (2005) suggested a relationship between childhood cancers and prenatal
proximity to emissions from large concentrations of industrial fossil fuel combustion in
Great Britain. In another study the malaria-like infection Human Babesiosis exhibited a
clumped distribution related to tick density, which may be used to predict areas of risk of
human infection (Mather et al. 1996).

Variation in attitudes toward a particular species in different locations has also
been explored. For example, Bowman (2001) surveyed landowners to determine whether
differences in attitudes toward black bears existed between areas of large (Arkansas) and
small bear populations (Mississippi). Neither demographic variables nor knowledge
about bears differed between the two locations. However, respondents in Arkansas, who
collectively have experienced more bear damage to crops, were less likely than
Mississippians to support increasing the bear population size. Comparative spatial
studies of attitudes have also been completed for wolf reintroductions (Pate et al., 1996),
grizzly bear restoration (Merrill et al. 1999), red panda protection (Fox et al. 1996), black
bear population expansion (Peyton et al. 2001), and cougar management (Riley and

Decker 2000b). Bowman et al. (2004) used a demographic model to predict support for

153

black bear restoration throughout Mississippi. Each of the foregoing studies has
suggested the importance of assessing attitudes in different locations. To date, however,
we are unaware of any studies that have applied spatial analysis of attitudes to assess
distribution of individual survey responses. Therefore, our goal is to assess the spatial
distribution of attitudes in the context of a potential black bear reintroduction.

The historical range of the Louisiana black bear (Ursus americanus luteolus)
included all of Louisiana, southern Mississippi, and southeastern Texas (BBCC 1997).
Widespread timber extraction and extensive hunting led to the near extinction of the
subspecies by the early 19003. By the second half of the 20th century, only two isolated
populations remained, both of which were in eastern Louisiana. Beginning in the 19905,
wildlife managers and bear conservation groups in Louisiana led an effort, through public
outreach, to build public support for bear recovery in Louisiana. The Black Bear
Conservation Committee and US Fish and Wildlife Service created recovery plans, an
objective of which was to restore the Louisiana black bear throughout its historical range
(Bowker and Jacobson 1995; BBCC 1997). To date, recovery in Louisiana has been
successful (Van Why 2003), and feasibility analyses were completed for Mississippi with
positive results (Shropshire 1996; Bowman 1999).

Although there is no known breeding population in eastern Texas, the number of
black bear sightings has increased during the past decade. These bears are likely
transients from bear populations in Louisiana, Arkansas, and/or Oklahoma. The number
of increased sightings prompted the creation of a bear management plan for both eastern
and western Texas by the Texas Parks and Wildlife Department. The ultimate goal of the

management plan is to restore habitat for the future reestablishment of black bear as a

154

viable ecosystem component in eastern Texas (TPWD 2005). Short-term objectives
(over the next 15 years) include public coordination, communication, outreach and
information dissemination, habitat management, and research (TPWD 2005).

Big Thicket National Preserve (BTNP) is targeted for initial steps for bear
recovery. BTNP was established as the nation’s first National Preserve in 1974 to protect
1 1 unique ecosystems found within its boundaries. Since establishment, BTNP has also
been designated as a UNESCO International Biosphere Reserve, an American Bird
Conservancy Globally Important Bird Area (IBA), and member of the United States Man
and Biosphere Program. The preserve’s 12 land and river corridor management units
together total 39,256 ha. Each preserve unit alone is not large enough to contain a viable
black bear population, so bears that live within the preserve are likely to wander beyond
the preserve’s borders. Because private land and residences sm'round the preserve units,
it is essential to determine if local residents will be tolerant of bear presence.

The objectives of this research were to: (I) ascertain the spatial distribution of
residents’ attitudes across the study area toward potential bear recovery strategies for the
southern portion of eastern Texas, and (2) determine whether distance from BTNP is a
factor in resident attitudes. We assessed support for three general management options
for black bears (described in methods section): (1) natural (non-assisted) recovery, (2)
human-assisted reintroduction (which included two equivalent strategies with slightly

different wording were assessed), and (3) no bear recovery.

155

6.2. Methodology
6. 2. 1 . Study area and survey design
Please refer to Sections 4.2.1 (including Figure 4.1) and 4.2.2 for descriptions of

the study area and survey design, respectively.

6. 2. 2. Dependent variables
The four dependent variables described and used in Chapter 5 (see Section 5.2.2)

analyses were applied to spatial analyses.

6. 2. 3. Non-response follow-up

Please refer to section 4.2.4 for discussion of the non-response follow-up survey.

6. 2. 4. Spatial analysis

We used DeLorme Street Atlas USA® 2004 (DeLorme, Yarmouth, Maine, USA)
to plot approximate locations of respondent’s addresses. Use of a GPS to identify
location of addresses, although more precise, was not a viable option for reasons of
respondent confidentiality. For addresses that were not available through use of
DeLorme Street Atlas USA®, a random point was selected from within the respondent’s
respective zip code using ArcView GIS 3.2 (Environmental Systems Research Institute,
Inc.). For spatial analysis, we converted respondent location information from Latitude-
Longitude to UTM XY coordinates.

To evaluate the spatial distribution, or clustering of responses, we used SaTScan

version 4.0 (Kulldorff et al. 1998). SaTScan software was designed to evaluate spatial

156

and temporal distribution of designated events (Kulldorff and Nagarwalla 1995; Kulldorff
1997). This software has been used to explore the spatial distribution of breast cancer
(Kulldorff et al. 1997), chronic wasting disease in deer (Joly et al. 2003), bacterial
infections of herbivores (Smith et al. 2000), West Nile virus (Mostashari et al. 2003),
learning disabilities in children (Margai and Henry 2003), and toxic parasites (Miller et
al. 2002). Because these examples all involve assessment of risk to a particular
population, we determined that SaTScan could also be used for spatial assessment of
human attitudes toward particular black bear population recovery strategies. This is
based on the assumption that residents who are unsupportive of a recovery program may
represent sources of greater risk to black bears if a recovery occurs (e. g., illegal killing).
The SaTScan’s spatial scan statistical procedure involved the creation of three
data input files: (1) a case file containing the unique identification number for each
respondent and corresponding binary variable on which a one represented a “yes”
response to a particular variable and zero equaled all other responses (i.e., no, unsure), (2)
a population file containing the unique identification number for each respondent and the
binary opposite to the case file (i.e., 0 = “yes” in the case file and all else = 1), which
serves as a control, and (3) a geographic file containing the X and Y coordinates for each
respondent. The scan statistic applied an infinite number of geographic circles to each
“yes” response location (observed responses) with the maximum circle size equal to half
of the total study area, and determined which circles contained a greater number of other
“yes” responses than spatial randomness (expected responses). In other words, by
applying geographic circles to each response, clusters of responses that were more likely

to occur than by chance alone were identified. This procedure was based on the

157

alternative hypothesis that there is a greater rate of occurrence of a particular event (in
our case a “yes” response) within a particular set of geographic circles than outside of the
circles. We applied a likelihood function for a Bernoulli model, calculated as:

(N-n)-(C-c)I-(x)

<c/n>°<1-c/n>‘"‘”<[C-c1/[N-n1>‘c‘°’<1-<[C-c1/[N-n1>
where C = the total number of “yes” responses (i.e., responses) across the study region, 0
= the number of “yes” responses within the particular geographic circle, N = the total

’9 ‘6

number of “yes” responses plus controls (i.e., “no, unsure”) within the data set, and n =
the total number of “yes” responses and controls in an identified cluster. Because our
objective was to identify high rates of a particular response, I(x) = 1 when the geographic
circle had more cases than expected for the null hypothesis (i.e., the rate of occurrence of
a given response was no different from random), and I(x) = 0 for all other instances. The
cluster least likely to have occurred by chance was identified as the main cluster, and all
other identified clusters were called secondary clusters. A maximum likelihood ratio test
statistic was calculated for all identified clusters. The p value for each maximum
likelihood test statistic and distribution compared to the null hypothesis was calculated by

repeating the same procedure under 999 random replications (SaTScan’s default number)

of the same data for the null hypothesis using Monte Carlo simulation.

6. 2. 5. Distance from the preserve
We used the Nearest Features extension (version 3.8) of ArcView to calculate the
distance from each respondent to the nearest boundary of BTNP. For each of the

dependent variables, we used a Pearson’s correlation to determine whether distance from

158

the preserve boundary was related to support (or lack thereof) for particular strategies.

All alpha values were defined at the 95% confidence level.

6.3. Results

Differences in support for each of the four potential recovery strategies are
discussed in Chapter 5 and displayed in Table 5.2. The spatial scan statistic procedure
identified 67 clusters among 14 responses (i.e., three variables with three possible
responses each; one variable with five possible responses; Table 6.1). Although a main
cluster was identified for each response for each variable, the number of secondary
clusters varied between variables. Several secondary clusters for a particular variable had
similar Log-likelihood ratios, indicating that the individual points were very close
together. There were two statistically significant clusters identified (Figure 6.2): (1) non-
support for Natural (Log-likelihood ratio = 9.88, p = 0.043), located in the north-central
portion of the study area in proximity to the Angelina National Forest, and (2) strongly
disagree for No bear (Log-likelihood ratio = 12.19, p = 0.005), located in the
southeastern, and suburban, portion of the study area.

There was a negative but non-significant relationship between distance from the
preserve and support for Natural (r = —0.022, p = 0.493), Assist (r = —0.018, p = 0.576),
and Restock (r = —0.048, p = 0.134). However, a weak significantly negative relationship
existed between distance from preserve and support for the No bear strategy (r = —0.065,
p = 0.045). In other words, respondents closer to the preserve were more likely to

support N0 bear.

159

 

6.4. Discussion

As noted above, short-term objectives of the Texas black bear plan include
extensive public outreach. Knowing the distribution of attitudes toward potential bear
strategies can provide guidance for managers when planning public information and
outreach sessions. For example, if a large number of residents located in proximity to
each other share antagonistic attitudes toward a reintroduction, bear managers may
concentrate outreach efforts on the location to better understand why such attitudes occur.

Few significant clusters of attitudes toward particular strategies were obvious.
The only statistically significant clustering that occurred here were non-support for
Natural (non-human assisted) and strong disagreement with N0 bear (excluding bears
from the area). Random chance and lack of a complete enumeration of all area residents
(or even substantially higher coverage) are potential reasons why only these two
clustering events occurred. Clustering of Natural also may be a result of: (1) the
individual respondents do not want bears in the area, or (2) the individual respondents
support more aggressive bear recovery strategies (i.e., reintroduction and restocking).
Both of these reasons are supported. Of the 11 respondents within the statistically
significant cluster, eight indicated support for either the Assist or Restock strategies, and
two indicated support for the No bear strategy (Morzillo, unpublished data). From the
same group, one respondent indicated “unsure” about both the Assist and Restock
strategies, but did not support N0 bear. For respondents who supported pro-bear
strategies, these results may suggest that residence near the national forest is related to an
interest in nature and wildlife. In fact, for other items in our survey, all eleven

respondents indicated an interest in wildlife (Morzillo, unpublished data), but only the

160

eight who supported a pro-bear strategy indicated that they would like to see bears in the
area. Of course, generalizing from these 11 individuals to all of the residents in the area
they represent is problematic because of their small number.

For No bear, the significant cluster of respondents who strongly disagreed (i.e.,
they did not support excluding bears altogether) was located within the most urbanized
part of our study area. Results from both Lohr et al. (1996) and Bowman (1999)
suggested that residents of urbanized communities typically have more positive attitudes
toward restoration of wolves and bears, respectively. Just the opposite, Peyton et al.
(2001) reported that residents in the heavily populated southern portion of Michigan
preferred limited bear presence even though the Michigan bear population is expanding
southward. However, results from our survey did not indicate that members of this
cluster were necessarily more likely to support bear recovery. Of the 24 respondents
within the N0 bear cluster, only five indicated support for Natural, 12 for Assist, and 17
for Restock. Between the three variables, there were also 20 “unsure” responses, which
suggests that even though a commonality existed for one management strategy among
individuals of these particular clusters, respondents were not in unanimous agreement
about an optimum strategy.

The limited number of statistically significant clusters related to particular
strategies implies that public outreach will likely be challenging. First, bear managers
will likely be met with a plethora of different opinions when outreach takes place in any
location, which makes generalizing results to a larger area difficult (e. g., census block
groups; Bowman et al. 2004), as well as customizing outreach programs for specific

locations. Second, movement of urban residents into rural areas may result in

161

amalgamation of urban and rural wildlife-related attitudes and values. Many residents,
particularly within the southern half of the study area, make long commutes to jobs in
cities such as Houston and Beaumont from rural communities. Attitudes of such
residents may be similar to attitudes of residents within larger communities (e. g., more
positive attitudes toward particular carnivores; Lohr et al. 1996; Bowman 1999) may be
increasingly predominant among rural residents.

The negative relationship between distance from the preserve and support for
Natural, Assist, and Restock was not significant, which further indicates a wide variety of
attitudes toward bear recovery exist within the study area. But respondents closer to
BTNP were significantly more likely to support the No bear strategy. This significant
relationship, although relatively weak, is critical information, especially because BTNP is
a target area for bear recovery. Bears that wander beyond the preserve’s borders may be
met with hostile actions from residents.

The suggestion that residents near BTNP are less supportive of a bear recovery
than residents further from the preserve may initially pose a big challenge to bear
recovery efforts. Bear managers must determine if conditions exist for which non-
supporters near the preserve will be willing to tolerate bear presence (e. g., movement of
bear if nuisance problems occur; financial incentives), or if such residents will
immediately threaten a bear’s well-being (e. g., shoot it immediately or use harmful
agents such as poisons). Should the bear population expand its range across the whole
study area, the possibility exists that attitudes across the study area will change. Having
residents throughout the study area who support a black bear recovery via Assist or

Restock may allow for widespread initial support for TPWD’s bear management goals.

I62

but support may decrease if bears become a nuisance at any time and in any given
location. Therefore, a large management challenge is to continuously monitor local
attitudes.

Unfortunately, several development proposals are threatening the ecologically
fragile landscape of BTNP and the surrounding area. Decisions involving divestment of
timberlands and urban development may drastically change the area’s landscape.
Rapidly growing cities in Texas, such as Houston, Austin, San Antonio, and Dallas-F t.
Worth, are playing a role in plans for an eight-lane superhighway and water diversion
projects. These threats have resulted in BTNP’s inclusion among the National Parks
Conservation Association’s (N PCA) ten most endangered parks in the United States
(NPCA, 2004). Land use decisions that threaten ecosystems are common in areas
targeted for species recovery programs. Spatial analysis of attitudes like that in the
current study can provide managers with insight into where local support and non-support

exist for conservation and species management goals.

163

 

$3 $4 cam 24 6 v m
8% Se 8; 2.: 86 a 2 3523 Emma.
33 N3. N3 :2 6 v m
83 a; 2: «S. a: 2 m
83 mi. ”3 EN 6 h m
EX. Rd 3% 3m 6 m m
33 an RN :1. a: S m
82 Re 3: o: _ :6 a 2 €253-85 fig
32 $6 a; E: 66 2 m
:3 mo.» 42 5.2 6 : mm 2 €256 ii
33 23 ta ”2 6 m m
82 3.0 ta 2: 6 o m
33 .43 0.2 OS: 36 3: 2 3523 =2:st
63 Ex worm m3 6 o m
was woo MB 8.3 66 3 m
82 m: a?“ $4 6 o m
E .o 8.» 3% 8.0 as t m
3.9.36 36 3m 22 2 c S 2 €o&a-§5 Benz
4de NS. 02 SN 6 m .m
336 N2. 3a SN 6 m .m
335 N2. 2% 8a 6 m .m
mad $4. 4: m; 2.6 t m
~85 $4 a; 3m 6 w m
$3 03 an as a: E m
was 2% 9% 3mm 6 o m
33 3% 3m ”.2 a: : 2 €256 Benz
6N3 corn—A503
03m>u& otmh woosznvxznwoxm xmt _mao< v—mt Umwooaxm nmommo no 59532 «.5530 ~33me “Gogowmcmz

 

.2365 snow :3QO 2: mafia mflwoabm EoEowmcmE 59502 .83 x83 E6228 @832 8353 60 $93.23. Ad 033.

164

nmwd
nmwd
hmwd
hmwd
vmcd
wamd
wad
mood
nmvd
hmvd
hmvd
oood
mead
god
ooad
mmwd
30.0
086
Sad
mvwd
mvwd
and
6&6
03.6
:56
Sad
mvod
momd
vomd

m_.m
m_.m
26
m_.m
wmb
mm.m
wm.m
om;u
cod
cod
cod
RN
5%
fimé
wmd
26
N_.m
Q6
Nmé
Rd
52m
36
$6
$6
$6
56
3.0
5:.
:8

mm.m
mm.m
mm.m
mm.m
and
3.3
3.3
3.2
nmdm
nmdm
nmdm
omw
NY:
56
mud—
w 6 .2
thN
3&3
:4“
Ext.
$.m
Sim
cod
5w;
SN
3N
VON
3N
NM:

$6
$6
de
de
Nth
omd
omd
m _ .c
3.0
ofio
omo
vmd
m fl .o
mod
2 .o
hmd
mod
36
mm.“
no;
no;
31V
mad
NOS
omd
$6M
NY?
5%
mmd

6 m
6 m
6 m
6 m
:6 t
6 N
6 N
6 N
6 N
6 N
6 N
6 N
6 N
:6 v
6 N
6 N
6 N
6 N
6 m
6 ...
6 6.
EV :
6 v
86 MN
66 o
6 N
6 N
a: S
a: :

2mmmEmmEmmmEmmmmmmEmmmmmEr/zmmm

6823 Son 02

62me Son 0 Z

68% Ewcobmv Son oZ

355:3 xooumom

9253505 xooumom

3.56:6 xoowmom

Afipcouv Nd 2an

165

_o>2 8:25:00 $3 05 3 oogocmcmm 880me AL xmtoamm 59
anew m 5.23 20:32 320 E3 mo :38 a 8m 8:86 8m 3593 5023 How magma—o $958?
News? 33333 a 55$» 3309.8 28%.. («o 59:82..

5520 58.80% H m roams—o 522 u E“

 

wood
cmod
«mad
cmod
v3.0
wood
owed
*mood
wmod
Nmmd
33.0
nmwd

emf
:.v
:d
Saw
:6
cow
cow
6.2
wad
mud
cod
Eh

mo.“
mm.m
mmxm
mm.m
mm.m
mm.m
mm.m
omd
on;
93m
moN
mm.m

mm.m_
g.—
um.—
cm;
vm;
mm;
mm;
$8

8.:
91m
3:.
de

66 NN
6 v
6 v
6 v
6 v
6 m
6 m

:6 2

G6 a
6 w

:6 S

6 m

Gouwmmzv bwcobmv 83 oz

Aoewmflwv Son 02

mmEmEmmmmmr/Jm

:Ccouv do 293

166

Figure 6.1. The 12-county region in southeastern Texas used to assess the spatial
distribution of resident attitudes toward potential black bear recovery strategies from
which survey participants were selected. These counties include and surround Big
Thicket National Preserve (shaded areas in circle).

167

 

 

 

Figure 6.1

 

 

 

 

San
Augustine f

 

 

Angelina

 

 

 

 

  
 
 

 

 

 

 

 

 

 

 

 

 

 

 

168

 

 

 

 

 

Figure 6.2. Statistically significant clusters of non-support responses for the Natural
(solid squares) and the strongly disagree responses for No bear (solid triangles) strategies
for black bear recovery. Also shown are the USDA proclamation boundaries for the
(clockwise from bottom lefi) Sam Houston, Davy Crockett, Angelina, and Sabine
National Forests (striped) and Big Thicket National Preserve (shaded).

169

CHAPTER 7

INTEGRATING ECOLOGY AND SOCIOECONOMICS FOR SPECIES

REINTRODUCTIONS

In collaboration with

J ianguo Liu

170

Abstract

To mitigate human impacts on wildlife species and their habitat, there are
increasing efforts to recover populations of locally extirpated species. Integration of
ecological and socioeconomic data is necessary for species recoveries targeted for
human-dominated landscapes. We developed a systems framework and model for
simulating how ecological and socioeconomic processes may affect feasibility of a
Louisiana black bear (Ursus americanus luteolus) reintroduction in and around Big
Thicket National Preserve (BTNP), Texas. Based on analysis of timberland and bear
habitat, as well as social survey data, model results suggested that a declining trend of
timberland area over time is likely to continue in the future. This has resulted in a
decrease in the area of bear habitat among timberlands over time, but the magnitudes of
habitat change have varied across counties. Although the gender ratio of each county is
projected to remain relatively constant, model results suggested increases in median age
across the study area, ranging from an additional 3.9 to 11.8 years, by 2020. From multi-
criteria analysis, Sabine, Trinity, and San Augustine Counties were identified as the most
ecologically feasible for bear recovery, and Newton, Jasper, and Tyler were the most
socially feasible. Areas with high ecological and social feasibility did not overlap.
National forests may be better potential bear release sites than BTNP, as their locations
correspond to the before mentioned counties. Public outreach aimed at increasing
residents’ knowledge about bears, as well as gaining insight into why residents do or do
not support bear recovery are necessary before deciding whether and where bears should
be released. The model framework provides a useful approach for assessing feasibility of

reintroductions may be adapted for conservation programs in other locations.

171

7.1. Introduction

Human activity is causing major changes to the earth’s ecosystems (Soule’ 1991;
Noble and Dirzo 1997; Vitousek et al. 1997; Chapin et a1. 2000), and these can be
detrimental to wildlife species and their habitat (Liu et al. 2001; Sillero-Zubiri and
Laurenson 2001; Parks and Harcourt 2002). To mitigate human impacts on wildlife
habitat there are increasing efforts to recover populations of locally extirpated species
(Reading and Clark 1996; Breitenmoser et al. 2001; Woodroffe 2001a). Consequently,
even if researchers determine that a reintroduction is ecologically feasible within a
particular location (e. g., adequate habitat area; van Manen 1991), this alone may not be
enough to ensure a successful reintroduction and species population establishment in
areas dominated by humans.

Particularly for large and widely wandering species, an additional challenge for
reintroduction programs is that species may wander into areas of human activity while in
search of new habitat, members of the opposite sex, or food. Wildlife does not recognize
political, management, or ownership boundaries (Soule’ and Sanjayan 1998; Liu and
Taylor), and movement may cause an individual to traverse land with a multitude of uses
and ownerships. When species wander into human-dominated areas, there is potential
uncertainty with respect to how humans will react to a species’ presence (Woodroffe and
Ginsberg 1998; Merrill et al. 1999). In such cases, all lands become important for
conservation (O’Connell and N033 1992).

Many researchers have studied either impacts of human activity on wildlife or
human attitudes toward wildlife, but rarely simultaneously. Understanding these

interactions is critical to better understand the intricacies of real world conservation

172

issues. Lack of integrating ecology and socioeconomics has led to many conservation
failures. For example, a great quantity of ecological research has been collected about
giant panda (A iluropoda melanoleuca) habitat. Until recently, socioeconomic research
was almost nonexistent. As a result, the Chinese government, even when providing free
housing, largely failed to move residents out of Wolong Nature Reserve (established in
1975 for panda protection). This resulted from lack of understanding two important
social needs: (1) elderly residents were hesitant to relocate and adapt to new lifestyles,
and (2) there was no available land in proximity to the free housing on which this
agriculture-based community could establish farms (Liu et al. 2001). Ability to integrate
ecology and socioeconomics will increase effectiveness of policy makers’ abilities to
create feasible and effective solutions to meet the needs of wildlife and promote human
welfare over both short and long terms (Holling and Meffe 1996; Liu et al. 1999).

Systems modeling offers a useful approach to integrate data from different
disciplines that are related to factors affecting wildlife and their conservation (Liu 1993;
Liu et al. 1995; An et al. 2001). Related to resource conservation, systems modeling
tools have been developed for analysis of wildlife populations (Liu 1993), household
level decision-making processes (An et al. 2001), and assessing the effects of different
management strategies on a particular resource (Liu et al. 1994; Liu et al. 1995; Turner et
al. 1995; McDonald et al. 2001).

The few systems models developed for wildlife reintroductions have focused
mainly on ecological factors that may influence reintroduction success. For example, van
Manen and Pelton (1997) developed a systems model to determine the probability of

reintroduction success for black bears in the Big South Fork area of Kentucky and

173

Tennessee. Similar to results of van Manen and Pelton (1997), other studies have used
models to assess how temporal changes in individual ecological factors may affect post-
reintroduction species populations (South et al. 2001; Steury and Murray 2004). A major
weakness of most modeling tools related to reintroductions is the lack of incorporation of
how human-influenced factors that affect reintroduction feasibility. Bowman et a1.
(2004) used results from a landowner survey, demographic data, and spatial distribution
of bear habitat to develop a spatially explicit model for predicting attitudes toward black
bear restoration across Mississippi. However, this study did not seek to project how
changes in demographic and ecological factors in the future may affect dynamics of
support for bear restoration.

We developed a framework for simulating how ecological and socioeconomic
processes may affect feasibility of a Louisiana black bear (Ursus americanus luteolus)
reintroduction in and around Big Thicket National Preserve (BTNP), Texas. The
Louisiana black bear historically existed throughout all of Louisiana, southeast Texas,
and southern Mississippi (Lariviére 2001). Overharvest and habitat destruction led to
near demise of this subspecies during the early 19008 (BBCC 1997). During the second
half of the 20th century, only two small remnant populations of the subspecies existed,
both in eastern Louisiana. An extensive public outreach campaign during the 19903 led
to recovery efforts for the subspecies in Louisiana, as well as the creation of two
management plans that focused on recovery throughout the historical range (BBCC 1997;
Bowker and Jacobsen 1995). Thus far, recovery in Louisiana has been successful (Van
Why 2003; Benson 2005) and feasibility analyses have been completed for Mississippi

(Shropshire 1996; Bowman 1999).

174

In eastern Texas, there have been an increased number of black bear sightings
during the past decade (N. Garner, Texas Parks and Wildlife Department (TPWD),
personal communication). Because no breeding population of bears is known to exist in
Texas, it is likely that these sightings are transient bears from Arkansas, Oklahoma, or
Louisiana that wander into Texas. Regardless, the sightings have prompted creation of
an East Texas black bear management plan. Objectives of the management plan for the
next 15 years include public coordination, communication, outreach and information
dissemination, habitat management, and research (TPWD 2005). The ultimate goal is to
restore habitat for the purpose of reestablishing black bear as a viable ecosystem
component in southeast Texas (TPWD 2005).

By integrating interdisciplinary data, this systems model is designed to lend
insight into the underlying mechanisms that may influence bear habitat over time.
Remote forests containing a variety of fruit and mast producing species are important
components of bear habitat (Pelton 1982; Lariviere 2001), and changes in the area and
distribution of forested land will affect habitat. A majority (>75%) of land in proximity
to BTNP is managed privately for timber production or publicly by the federal
government (i.e., BTNP and four national forests). In addition, there is a local population
of human residents. Therefore, land use affecting bear habitat, as well as human
tolerance of bear presence, will be important factors that drive bear habitat dynamics.

Our first objective was to develop a systems model that incorporated bear habitat,
land ownership, and socioeconomic variables affecting attitudes toward black bears in
southeast Texas. Ideally, locations with a greater amount of habitat will also be more

socially feasible and support a reintroduction. Much of the bear habitat in southeastern

175

Texas is contained within public land and private timberland in rural areas. If more
highly suitable bear habitat (Pelton 1982; BBCC 1997; Lariviere 2001; see Chapter 2) is
available within a particular location, there may be less chance that bears will wander and
come into contact with human activity. We evaluated our model at two future time
points: (1) 10 years into the future (10 years after model parameterization, or 2010), and
(2) 20 years into the future (20 years after model parameterization, or 2020). The year
2020 coincides with the end of the lS-year term of the present black bear management
plan (TWPD 2005). Our second objective, based on results of the first objective, was to
rank the counties within the study area in regard to both ecological and social feasibility
for the planned black bear recovery. We next describe the study area, model structure
and methodology used for parameterization and testing of the systems model, present the
results of model implementation, and discuss the implications of the results for bear

reintroduction management.

7 .2. Methods
7. 2. 1. Study area

Our study area consisted of twelve counties in southeast Texas, which includes
the 39,28S-ha Big Thicket National Preserve (BTNP; Lat/Long. —94.36/30.3 8) and four
National Forests (Angelina, Davy Crockett, Sam Houston, and Sabine; Figure 1). BTNP
is unique in that it consists of 12 disj unct management units, seven of which are river
corridors that connect larger non-corridor units. Approximately 500,000 residents live
within the study area (US Census Bureau, Census 2000 data). Much of the area is rural,

with numerous small towns throughout and one larger community (City of Lufldn) along

176

the northwestern boundary. The southern edge of the study area, which is also the most
densely populated portion of the landscape, consists of suburban development from

Houston and Beaumont.

7. 2. 2. Model design

We adapted three principles, as defined and described by Xie et al. (1999) for
model design. First, we sought to design a simple model that could easily be reproduced
for other species in other locations. A large portion of the raw data may be easily
collected from government agencies (e.g., land cover images), or is publicly available on
the Internet (e. g., census data). Although counties may be divided into sub-counties and
census blocks, we focused this study at the county level because this is the smallest
standard political unit for Texas. Unlike may other states in the midwestem and western
United States, the Texas cadastre does not follow the Public Land Survey System of 1785
(i.e., township and range system set from a principal meridian), which makes divisions
beyond the county level (although they do exist) inconsistent and complicated.

Second, the model must predict feasibility as consistently as possible across the
landscape, which will be important for minimizing bear-human conflicts. If certain
locations do not contain adequate bear habitat, or contain residents with negative attitudes
toward black bears, then managers may want to closely monitor bears that unknowingly
wander into these areas. Conversely, if particular locations contain a greater amount of
habitat and positive attitudes toward bears, these may be better locations for bear release

and management.

177

Finally, similar to objectives of Xie et al. (1999), this model was developed for
the purpose of improving bear management and policymaking. We sought to create a
user-friendly model that can be used by individuals with minimum modeling experience.
A graphical user interface is useful for testing and simulating how changes in each
variable may affect a preferred bear management scenario for a particular county. A
major assumption of the model is that the human and natural landscapes are dynamic,
meaning that the presence of habitat, particular landowners, and attitudes change over
time. As new data become available, as well as demographic and ecological
characteristics of the region change, bear managers may add data to the model to project
how changes may affect local support for bears in the area. The Texas Parks and
Wildlife Department may also use the systems model to refine management goals, plan
public information forums, and ascertain the potential for bear reintroduction and

management over the long term.

7. 2. 3. Model structure and quantification

The conceptual framework for this research is illustrated in Figure 7.2. Our
model consisted of four components: (1) Timberland, (2) Habitat, and (3) Social, and (4)
Multi-criteria evaluation. The timber component (see Section 7.2.3.2) directly affects the
habitat component (see Section 7.2.3.3.). The location of and harvest strategies used for
timber management directly affect the quantity, quality, and distribution of bear habitat.
For example, heavily managed fast-growing tree species (e.g., pine plantation) may be
favored for timber production. However, clearcuts can result in an extremely rigid

boundary and may provide limited cover until vegetation fills in the area. As a result,

178

bear managers may prefer selective harvest of particular size stands and enhanced
management for hardwood mast-producing species (Weaver et al. 1990). Important
characteristics of black bear habitat include quantity (i.e., area), quality (i.e., highly
suitable versus marginal), and distribution (Pelton 1982; Bowker and Jacobson 1995;
BBCC 1997; Liu et al. 1999; Lariviere 2001).

The social component (see Section 7.2.3.4) was developed as a separate model
entity. Human support for bear presence may be affected by personal attitudes toward
black bear, which may result from knowledge (or lack thereof) about bear, or concern
about how bear management policies may affect land use by local residents. For
example, federal mandates for bear habitat conservation may prohibit landowners from
particular activities on their lands if it is known that a mother bear and her cubs are
denning nearby. Such limitations on human activities may result in negative opinions
about bears and bear managers. The data from the timber, habitat, and social component
of the model were entered into a multicriteria analysis (see Section 7.2.3.5), used to
spatially assess and rank the ecological and social feasibility of a reintroduction across all
counties in the study area. We completed a sensitivity analysis (Haefner 1996) on all
independent variables to determine how changes in median age and percent male values
affect model results.

Stella (Higher Performance Systems, 1997) is a multi-level, hierarchical
environment for model development and implementation that allowed us to create a
simple model for management purposes. The model operated on a yearly cycle (i.e., one

time step = one year).

179

7.2.3 .1 . Land classification

We designed a simple method of land classification to emphasize factors that will
likely be influential in bear habitat dynamics. We divided land within the study area into
three categories: (1) public land, (2) private timberland, and (3) other. We created
several assumptions related to land ownership to simplify model creation. First, for
public land, we assumed that changes in the total area of public land would be minimal
over time and public land would remain under the current jurisdiction indefinitely.
Second, although >90% of land in BTNP and >75% of National Forest land was
classified as highly suitable (see Chapter 2 and 3), we assumed that the area of highly
suitable bear habitat among public lands would remain consistent over time. Third,
because bears are adaptable to many different habitats but prefer refuge from human
activity (Pelton 1982). Therefore, we assumed that the permanent designation of public
land as such be of greater importance to long term maintenance of bear management than
minimal changes in habitat among timberlands, which could be sold and converted to a
different land use at any time. Fourth, we concentrated modeling efforts on lands within
the public and private timberland categories as they are of most interest to bear managers.
Therefore, we assumed that changes within the remaining approximately 20% of “other”
land within the study area was, at this time, of minimal interest to bear managers
compared to changes in timberlands. As a result of these assumptions, we focused our

model construction on private timberlands.

7.2.3.2. Timberland component

The objective of the timberland component was to simulate changes in the area of

180

privately owned timberland over time. We collected timberland data from the early
19703, 1987, 1990, and 2002 from available land ownership maps (Acme Map Company,
Tyler, Texas), local timber landowners, and the USDA Forest Inventory and Analysis
National Program database (USDA Forest Service, Arlington, Virginia, USA). For each
time period, we determined the area of land owned by private timber companies in each
county (see also Chapter 3). Based on the difference in timberland area within each
county from one time period to the next (e.g., area of timberland in 1987 and then in
1990), an estimated percent rate of change of timberland area for each county was
calculated. For four time periods, this resulted in three different rates of change for each
county. We assumed that sale of timberland (i.e., a decrease in timberland area) would
take place only if selling would be economically beneficial to the landowner.

Several private timber landowners own land in more than one county, as
determined from land ownership records. Therefore, rates of timber gain or loss occurred
at the level of the study area. Also assuming that any timberland lost was no longer
timberland, we assessed all rates of timberland change acrossall counties, which ranged
from a net loss of —0.057% to a net gain of 0.026% per year. Plotting these rates on a
graph produced a normally distributed probability function (one-sample Kolmogrov-
Smimov Z = 1.186, P = 0.120; Figure 7.3). Using the timberland area of each county as
the baseline (t = 0) value, a random value from the range of the probability function was
drawn to simulate the percentage of timberland gained or lost for each time interval
within each county. If the random number was less than the mean value of the
probability function, habitat was lost for the given time interval. If the random number

was greater than the mean value of the probability function, habitat was gained for the

181

given time interval. By multiplying the area of timberland at time t by the randomly
generated rate of change value for the given time interval, we determined the change in
area of timberland from t to t + 1. By adding (or subtracting) the area of timberland
calculated at each time interval to the total area at time t, we calculated the estimated
changed and new total timberland area fort + 1. We repeated this process for each

county. Figure 7.4 illustrates the model structure of the timberland component.

7.2.3.3. Habitat component

The objective of the habitat component was to simulate changes in highly
suitable, suitable, marginal, and unsuitable habitat among timberlands over time. Using
ERDAS Imagine 8.7 (Leica Geosystems GIS & Mapping, LLC, St. Gallen, Switzerland),
we classified land cover data from the mid-19703, mid-1980s, mid-19903, and 2002
based on black bear ecology for the southern United States (Pelton 1982; Wagner 1990;
Weaver et al. 1990; Marchinton 1995; Maehr et al. 2003; Larkin et al. 2004; see also
Chapters 2 and 3). For each time period, we determined the proportion of area of each
county’s timberland that contained habitat of each of the four habitat categories.

To integrate the habitat component with the timberland component and spatially
allocate results of changes in timberland across the study area, the change in each
county’s timberland area (gain or loss) for each time interval from the model’s
timberland component was used to quantify simulated changes in bear habitat among
timberlands (Figure 7.5). The total amount of timberland gained and lost in each time
interval was used to determine the total area of habitat in each county that was affected

within the same time interval. If the amount of timberland change in the timberland

182

component was positive, then that amount of timberland was proportioned among and
added to each of the four habitat classes within the habitat component. If the amount of
timberland change in the timberland component was negative, then that amount of
timberland was proportionally divided among and subtracted from each the four habitat
classes. Because the model does not specify the detailed locations of changing habitat,
we assumed that the amount of habitat within each class that was gained or lost in each
time interval (i.e., proportionally added to or subtracted from each habitat class) followed
the current pattern of habitat suitability for each county. For example, timberland within
Angelina County was 62.4% highly suitable habitat. If the timberland component
produced a gain of 1,000 ha of timberland within Angelina County for a particular time
interval, then the total amount of highly suitable habitat within Angelina County would
increase by 624 ha (62.4% of the total timberland gained) as a result of the increase of

timberland.

7.2.3.4. Social component

The objective of the social component of the model was to assess how changes in
particular socioeconomic variables are influential as factors in predicting the feasibility of
a black bear recovery across different counties (Figure 7.6). During J anuary-March
2004, we surveyed 3,000 local residents to assess attitudes toward black bears and local
support for a black bear recovery. With a response rate of 40% after removing bad
addresses and names of those who did not wish to participate (n = 1,006), we applied
ordinary least squares regression (Babbie 1990, Sokal and Rohlf 1995) to determine

particular socioeconomic variables that affected attitudes toward black bears and a bear

183

population increase (see Chapter 4), preferred bear recovery management scenarios (see
Chapter 5), and spatial distribution of attitudes toward bears (see Chapter 6). We focused
only on variables that affected attitudes toward black bears for model parameterization
based on two assumptions. First, because the goal of this research is to develop a
framework for determining the feasibility of a black bear reintroduction, we assumed that
a reintroduction would take place in the future. Second, residents with positive attitudes
toward bears are more likely than those with negative attitudes to support a reintroduction
in the future, although this is not guaranteed (Lohr et al. 1996). Therefore, we assumed
that attempting to foster positive attitudes toward and increasing knowledge about bears
is an important first step before determining a specific recovery management method.
From our survey, six variables used to measure attitudes toward black bears were
statistically significant (Table 7.1): (1) gender, (2) age, (3) participation in passive-
appreciative activities related to wildlife, (4) duration of local residence, (5) knowledge
about bears, and (6) whether the respondent has seen a bear in the wild. We then
developed criteria for incorporating variable values into the model. Because our survey
data were not representative at the county level, use of survey data for model
parameterization of gender and age variables (for which better data are available) would
have resulted in a misleading product. Therefore, we used more—reliable US Census
Bureau data for model parameterization. Some historical data for gender and age were
available from the US Census Bureau. Because males were more likely than females to
have positive attitudes toward black bears (based on our survey data), we determined the
percent of residents that were male within each county from census data, for which such

information was available from 1990 and 2000. Because only two data points were

184

available, we calculated the average percent of residents who were male explicitly based
on these two data values. For example, in Angelina County, 47.3% of residents were
males in 1990. In 2002, 48.2% of residents were males. This resulted in a 0.019% rate
of change in the percent of residents who were male between 1990 and 2000 (a lO-year
period), or a 0.0019% annual rate of change. The 0.0019% rate of changed was used in
the as the simulated rate of change in percent males in Angelina County over time. We
repeated this calculation using census data for each county, and used the historical rate of
change to project how each county’s percent male residents would change in the future.
We also based model parameterization for age on data available from the US
Census Bureau. Because younger residents were more likely to have positive attitudes
toward black bears (based on survey data), we assessed the median age of county
residents from census data, available for 1980, 1990, and 2000. With three data points
available, we calculated the average rate of change, based on census data, in median age
over time. For example, the median age of Angelina County residents was 28.9 in 1980,
32.3 in 1990, and 34.2 in 2002. This resulted in an average rate of change of 0.088% per
decade, or 0.0088% per year. We repeated this calculation using census data for each
county, and used the historical rate of change to project how each county’s median age
would change in the future. In a sensitivity analysis, we varied the values by 10% to
determine how changes in median age and percent male values affected model results.
For the remaining four variables (participation in passive-appreciative activities,
duration of residence, knowledge about bears, and see bear in wild), no historical data
were available for model parameterization. The only data available were those collected

by survey implementation. Therefore, we assumed that the value of each of these

185

variables was static and equal to the average value of survey responses at the county
level. For example, the average knowledge value for Angelina County respondents was
2.7 bear knowledge questions answered correctly out of a possible six. For model
parameterization, we assumed that the value of 2.7 was constant, but varied variable
values by 10% in sensitivity analysis to determine how changes in each variable affected
overall model response. This process was repeated for the remaining three variables and

for each county.

7.2.3.5. Multi-criteria evaluation

A multi-criteria evaluation (Voogd 1983) was used to spatially assess how
individual counties rank in terms of favorability for bear recovery strategies. In other
words. this component determined whether feasibility of a bear reintroduction, based on
the variables used, was greater or lesser in any one county than in others. Parameters
(based on variables mentioned above) may be varied, and additional variables may be
added (or present variables removed) to reflect potential management activities or
policies, and how the activities and policies affect and are affected by variables of
interest.

The multi-criteria evaluation was divided into two components: ecological
feasibility and social feasibility. This was based on the rationale that some counties may
rank high ecologically, but low socially, or vice versa. The ecological component
contained four variables quantified at the county level: area of highly suitable habitat
(ha), area of suitable habitat (ha), area of marginal habitat (ha), and are of public land

(ha). We did not include area of timberland because this was the variable from which we

186

derived habitat and including it would be redundant. Area of unsuitable habitat (ha) was
excluded because, although we quantified its presence among timberlands and it could
have a negative impact on a reintroduction, its presence is of minimal importance
compared to the other habitat classes (i.e., as discussed below, unsuitable habitat would
have been assigned a weight of zero). In other words, total area of highly suitable,
suitable, and marginal habitat are more critical for estimating the number of bears that
could live within a county. The social component consisted of the six social variables
described in Section 7.2.3.4. Age and gender information were determined from Census
2000 data, and data for the remaining four variables were determined from the social
survey. We evaluated county rankings for both multi-criteria components under three
scenarios: ( 1) present, (2) 2010 (projected using model simulation), and (3) 2020 (also
projected).

Variables within the procedure were assigned fixed weights (sum of weights = 1)
based on the relative importance of each variable as related to feasibility. For example,
of utmost ecological importance for bears are highly suitable habitat and public lands,
which both may serve as refuges (see also Chapter 2). Therefore, we assigned the largest
weights to highly suitable habitat and public land variables (weight = 0.35 each;
cumulative = 0.7). These equal rates are the result of a tradeoff, such that large areas of
highly suitable habitat exist but a large portion of it is among private lands that are not
indefinitely guaranteed to exist (i.e., could be sold and used for urban development). On
the other hand, public lands may not contain as much highly suitable habitat collectively,
but they are more likely to exist indefinitely. Suitable habitat is likely to be used by bears

(for foraging) and regularly traversed, but not likely as important as highly suitable

187

habitat (weight = 0.20; cumulative 0.2 + 0.7 (from weights of highly suitable habitat and
public land) = 0.9). Marginal habitat, which may be traversed but is not likely to provide
long-term refuge for bears, is likely of least importance (weight = 0.10, cumulative = 1).
We calculated the county rankings for each of 200 simulations to determine the
frequency of individual ecological ranking for each county. To test the consistency of the
results, we also calculated and compared the county rankings for the first 100 versus
second 100 siulations.

Social variables were based on the magnitudes of standardized coefficients as
determined from regression as a proportion of the total of all coefficient values summed
(see also Chapter 4). In order of descending magnitudes of influence, the variables and
resulting weights were: knowledge (0.205), participation in passive-appreciative activities
(0.198), gender (0.193), age (0.162), length of residence (0.141), and whether the
respondent has seen a bear in wild (0.101). The value of each variable was multiplied by
the respective weight and summed across variables for each county to produce
standardized ranking of counties a scale of 0-100 (Voogd 1983).

Comparing ecological and social components will allow managers to weigh
management options based on ecological and social feasibility. For example, if Angelina
County contained a high ecological ranking (e. g., large area of highly suitable habitat) but
a low social ranking (e.g., low scores for questions about bear knowledge), managers will
know that habitat is present, but extensive public outreach will be necessary to determine
if bear reintroduction into Angelina County is truly feasible. Just the opposite, if
Angelina County contained a low ecological but high social ranking, managers may focus

public outreach on resident land use management for improving bear habitat.

188

7.2.4. Simulation scenarios
We ran the baseline simulation (i.e., timberland and habitat occurrence based on

current land cover and land use information) 200 times to estimate projected changes in
timberland and four classes of bear habitat (all four classes). After 100 simulations, the
standard error was less than one percent of the mean for each simulated variable. We re-
ran the model a second 100 times to confirm the statistical relationship between the two
sets of simulations. To do this, we constructed a 95% confidence interval around the
mean of all 200 simulations (sum of all runs from both sets of simulations). The mean of
each set of 100 simulations was statistically similar to the pooled 200 simulations in that
the 95% confidence interval of the 200 simulations contained the individual means of
each set of 100 simulations.

To verify the model, we simulated timberland and habitat area for the present
(2002) by using 19703 data as the baseline (i.e., projecting to the present from 19703).
We statistically compared the model predictions to the observed values of timberland
area and area of each of the four habitat classes. To do so, we constructed a 95%
confidence interval around the mean of the simulated model predictions. The observed
versus the predicted model values were deemed to be significantly different if the 95%
confidence interval around the mean predicted value did not contain the observed value.

We also sought to estimate how timberland and habitat may change in the future.
Therefore, simulation results were projected to year 2010 and 2020. These time periods
(year 2010 and 2020) corresponded to (1) 10 years after initial data collection and five
years after initial implementation of black bear plan, (2) 20 years after data collection and

15 years after initial bear plan implementation, respectively. The year 2020 also

189

corresponded to the end of the time period for which the 15-year black bear plan’s
present form will expire (TPWD 2005). We ran baseline simulations for each variable
for model verification, as well as to estimate simulated median age and percent male
values for 2010 and 2020.

To assess how sensitive the overall ecological and social county rankings were to
changes in individual variables (sensitivity analysis; Haefner 1996), weights that were
used for multi-criteria evaluation were increased and decreased 10% for each variable.
Because management activities would focus on particular variables, such as increasing
residents’ knowledge about bears, a 10% change in standardized variable values was also
tested. However, changing the value of each county or any combination of n counties
would result in an exceptionally large number of changes to be made with likely little
effect on the overall interpretation of the model output. For example, we could increase
the standardized value for knowledge in Angelina County alone or change the knowledge
values of Angelina and Hardin Counties together, or change Angelina, Hardin, and Tyler,
etc. Because bear managers are most interested in the most and least feasible locations
for a recovery, standardized parameters were adjusted for the highest and lowest-ranked
counties to determine if the new values would result in a significant change in rankings.
Starting with a 10% change, we adjusted the weights and standardized variable values

increasingly larger percentages until the changes resulted in a difference in rankings.

7 .3. Results
For model verification, we used calculated rates of change to project present

conditions based on past data. For timberland, the model underpredicted observed

190

timberland area in Angelina, Jasper, Newton, Orange, Sabine, San Augustine, San J acinto
and Trinity Counties for 2002 (Figure 7.7). Timberland area was overpredicted for
Hardin, Polk, and Tyler Counties.

For verification of highly suitable habitat area, highly suitable habitat was
underpredicted all counties except Tyler, which was underpredicted (Figure 7.8).
Similarly, no statistically similar observed and predicted habitat values existed for
suitable, marginal, or unsuitable habitat. Suitable habitat was underpredicted for all
counties except Tyler (Figure 7.9). Marginal habitat was overpredicted in Angelina,
Hardin, Liberty, Polk, San Jacinto, Trinity, and Tyler Counties (Figure 7.10), but
underpredicted in Jasper, Newton, Sabine, and San Augustine Counties. Unsuitable
habitat was underpredicted for all counties except Polk and Tyler (Figure 7.11). The
magnitude of habitat overprediction or underprediction varied across counties and habitat
classes. For future projections of timberland and habitat, the models for timberland and
all habitat classes suggested a decreasing trend across all counties (Figure 7.12; Figure
7.13; Figure 7.14; Figure 7.15; Figure 7.16), following the decreasing trend in timberland
area since the 19703. Because the observed timberland and habitat values were within
95% confidence limits of only one projected value each (Trinity timberland, Polk highly
suitable), we were unable to validate timberland and habitat data.

For the social component, verified median age data were minimally different from
observed median age (Table 7.2). The only differences between observed and verified
values were because of rounding. Because verified results for median age were based on
a simple mathematical operation using a consistent rate of change in median age between

time periods. we were not able to statistically test for significant differences between

191

actual and verified data or report standard errors. Simulated future changes in median
age over time suggested changes in median age, on a county basis, ranging from 2.0 to
5.8 years by 2010, and 3.9 to 11.8 years by 2020.

Verified value matched observed value for Trinity County. However, the model
overpredicted the observed male percent of each county’s population for nine of the 12
counties (Angelina, Jasper, Newton, Orange, Polk, Sabine, San Augustine, San Jacinto,
and Tyler Counties; Table 7.3). Two counties (Hardin and Liberty) were underpredicted,
both by approximately 1.5%. Similar to percent male, because verified results were
based on a simple mathematical operation using a consistent rate of change in percent
male between time periods, we were not able to statistically test for significant
differences between actual and verified data or report standard errors. The greatest
magnitude of change over the next 20 years in percent male residents of a county is
projected for Trinity County (change of 15.8 percent). Three counties were predicted to
have a loss in the percentage of males (Jasper, San Augustine, and San Jacinto), whereas
the remaining nine counties were predicted to have a gain in the percentage of male
residents.

Table 7.4 shows the frequency of ecological rankings for counties within the
study area. Sabine and Trinity Counties were consistently ranked as the counties where
ecological feasibility of a reintroduction was the greatest, whereas Liberty and Orange
Counties were most consistently ranked as the lowest. Projected rankings for 2020
contained less variability in individual county ranks than for 2010. Newton, Jasper, and
Tyler Counties ranked as the three counties with the greatest social feasibility (Table 7.5)

at present and as projected for the future. Trinity, Orange, and Polk were the least

192

socially feasible counties. Results of rank sum results suggested that no counties ranked
as both high ecological and social feasibility (Figure 7.17). Regardless of whether
individual counties contain a large area of highly suitable habitat among timberlands

(e. g., Jasper or Newton; Figure 7.17), a relatively small amount of public land within a
county would result in the county having a lower ecological ranking.

For sensitivity analyses, both the weights and standardized values from multi-
criteria analysis were adjusted to determine how changes in both would affect the overall
model. For ecological feasibility and with a 10% change in weights, increasing the
weight of highly suitable habitat was the only change that resulted in a different ranking
of counties, and with only minimal changes. San Augustine and Hardin Counties
switched from 3 and 4 to 4 and 3, respectively. Liberty and Newton Counties also
swapped rankings between 10 and 11. In fact, adjusting the weights for highly suitable,
suitable, and marginal habitat classes by any amount of value did not result in changes to
the top and bottom-ranked counties. Decreasing the weight of public land area (while
increasing the weights of the other three variables in equal amounts to account for the
decreased weight of public land) resulted in a change in variable rankings. The weight
for public land area had to be decreased from 0.35 to 0.2 for Sabine County to be
replaced by Trinity County as the most ecologically feasible, but Orange County did not
move from least ecologically feasible with any change in weight. Weight of public land
area also had to be decreased from 0.35 to 0.2 to result in a similar switch between
Sabine and Trinity Counties for 2010 and 2020. Therefore, the habitat van'ables were
robust to changes in weights in sensitivity analyses, but changes to public land area had

the most significant effects on model’s overall results.

193

Highest and lowest-ranked counties were also robust to 10% changes in the
standardized values for variables associated with ecological feasibility. A 15% change in
the standardized value of area of public land was needed for Trinity County to become
the top-ranked county ecologically, but Orange County remained the lowest regarless of
the change in public land. Even a 100% increase in the standardized values of habitat
variables did not change the overall rank of the 12 counties.

For social feasibility, county rankings were robust to a 10% change in both
weights and standardized values. Newton and Jasper Counties ranked as first and second
socially, respectively, for all attempted variations in weights. Decreasing the weight of
knowledge to 0.1 (a decrease of approximately 50%) did not result in a difference in the
socially top ranked county, but Tyler replaced Jasper County as second. However, the
county rankings were robust to changes in the remaining social variables. Social rank of
Trinity County remained last regardless of changes in weights.

A 50% increase in the standardized value of knowledge was needed to replace
Newton County with Jasper County as the socially top-ranked county. Similar to
sensitivity of weights, Trinity County also remained last-ranked for all changes in
standardized values. Changes in county rankings for 2010 and 2020 projections based on

altering standardized values were consistent with changes in rankings for the present.

7.4. Discussion and conclusions
For this study, we attempted to create a simple systems simulation model to assess
feasibility of a black bear reintroduction across our study area. Although our model

needs to be improved in the future, it provides a creative framework that bear managers

194

in Texas, as well as managers involved with other species reintroduction and wildlife
conservation efforts, may use to assess how both ecological and socioeconomic variables
may affect feasibility of a reintroduction.

Ideally for bear recovery, counties with the highest-ranking ecological feasibility
would also contain the greatest social feasibility. In general, this was not the case (Table
7.4; Table 7.5). Based on the results of this model, counties with high ecological
feasibility were moderate or low socially, and vice versa (Figure 7.17). Sabine, Trinity,
and San Augustine Counties contained a large amount of national forest land, as well as
timberlands. In fact, these counties, as well as Angelina, Hardin, Newton, Jasper, and
San J acinto Counties, also contained large areas of core habitat (see Chapter 2).
However, although Jasper and Newton Counties contained a large amount of highly
suitable habitat (Figure 7.17), the small amount of public land within the same counties
was highly influential in these counties receiving a low ecological ranking. Therefore,
management of private land will be critical for maintaining bear habitat in Jasper and
Newton Counties over time. For counties that contained a large amount of public land
(e.g., San J acinto, Sabine, and San Augustine), maintenance of highly suitable habitat is
not as dependent on private land management.

Low social feasibility for high ecologically ranking counties may suggest that
management efforts in these counties will be mainly people-related to ensure a healthy
bear population. San J acinto, for example, contained a large amount of national forest
land although its ecological feasibility was moderate. A reason for this may be that
remaining land within this county was not largely timberlands. However, with a

moderate ecological feasibility and high social feasibility, management efforts focusing

195

on land use to enhance black bear habitat in San Jacinto County may have fast and
positive results. Therefore, across the study area, tradeoffs between ecological or social
management foci must be made and customized for each county.

In counties that rank high ecologically, more emphasis must be placed on social
outreach agendas such as focus groups and information forums. As suggested by
sensitivity analyses, knowledge about bears was the social variable to which changes in
the overall rankings were most sensitive. In addition, knowledge is the most likely the
variable that bear managers may exert influence (i.e., bear managers are not likely to be
able to change the gender distribution of a county, or how often residents observe
wildlife). Therefore, efforts to increase residents’ knowledge about bears may increase
overall social feasibility across counties. Just the opposite, socially high-ranking counties
may seek guidance about land management for bear habitat. Because public land area
was the most influential variable in ecological sensitivity analyses, bear managers may
seek to encourage interested residents (or local corporations) to set aside land in
conservation trusts or easements to promote bear habitat conservation. Likewise,
donations of land to organizations related to land management (e.g., The Nature
Conservancy) might also assist in the indefinite conservation of land as bear habitat.

Counties that contain BTNP ranked moderate ecologically, and high to moderate
socially. The moderate ecological ranking is logical, as the twelve BTNP units only
collectively total an area of 39,000 ha, and Tyler County has the potential to lose a large
amount of forest via divestment of most of its timberlands in 2002 (see Chapter 3). The
high to moderate social ranking contradicts results found in Chapter 6, specifically that

residents closer to the preserve (compared to further away) were less supportive of a

196

black bear reintroduction. Because of this uncertainty, if BTNP is to remain a target for
bear release, more intense research into local support for a reintroduction, as well as more
detailed study of nearby habitat are necessary. Based on the results from this study,
national forest lands may be a better and more feasible alternative for bear recovery.
However, verification of attitudes toward bears in proximity to Angelina National Forest
is also needed because this area was identified as a potentially supportive location for
bear recovery based on spatial assessment of attitudes (see Chapter 6). Consequently,
San Augustine County did not rank high for social feasibility although Jasper County, the
northern portion of which is within Angelina National Forest, did rank high for social
feasibility. Strategically, the location of these national forests along the northern portion
of the study area may provide refuge for bears moving between populations in Oklahoma,
Arkansas, and Louisiana.

We suggest two main reasons why additional data are necessary for improving
verification of timberland and habitat. First, the change in timberland area from 1990 to
2002, and the resulting annual rate of change, take into account the total divestment of
one timber company’s lands (approximately 10% of all timberlands; see Chapter 3) in
2002. It is possible that the resulting rates for simulating change in timberland area were
greatly affected by this atypical one-time loss. More specifically, the sharp rate of
change from 1990 to 2002 may have resulted in unnaturally high rates of change in the
simulation probability distribution. This may lend insight into the underprediction of
timberland area in several counties. However, Tyler and Newton Counties were most
affected by the timberland divestment (see Chapter 3) and timberland was overpredicted

for Tyler yet underpredicted for Newton in model simulation (Figure 7.7). With only

197

three years worth of data for assessing ownership, changes within large time gaps may
have went unrecorded. The large divestment of 2002 offers great opportunity to track in
detail changes in land ownership, and how ownership changes affect changes in habitat.

Second, modeling efforts were focused on timberlands because of their potential
to maintain large areas of bear habitat over time. Although this focus, combined with
consideration of assumptions for public lands, allows us to encapsulate changes across
approximately three-quarters of the study area, future addition and analysis of changes to
the “other” land ownership category will further understanding of ownership and land
cover changes across the study area. Adding land cover data for intermediate years may
allow for identification of former timberlands that are now within the “other” category,
and track both rates of change as well as before-and-after land cover characteristics in
such locations. Expansion of human activity in the form of residential and urban
development into timberland areas will likely be driven by precisely what timberlands are
sold or not sold. In other words, presence of timberlands will constrain land use change
to locations within the “other” land classification category. Therefore, the rate of change
of land use within “other” locations may be greater than the general rate of land use
change across the study area. Therefore, incorporation of “other” land into the model as
its own component will likely produce the most accurate results.

Improvements may also be made to the social component of the model. First, the
survey data used here were collected from only a small portion of area residents. Specific
limitations to survey data (see also Chapter 4 and 5) included an underrepresentation of
females and minority ethnic groups, as well as data collection from specific stakeholder

groups such as hunters and conservation groups. Our survey respondents were

198

disproportionably young compared to the actual population reported in the census data.
Only further data collection will provide better insight into resident attitudes. Second,
this model did not incorporate demographic variables such as population and number of
households. As in other locations within the US and worldwide (Liu et al. 2003), the
human population of southeastern Texas has increased dramatically over the past 30
years, but the number of households has increased even more rapidly, particularly along
the southern and western portions of this study’s area (US Census Bureau data).
However, growth is limited to the land that was classified as “other” in this model.
Therefore, by adding the “other” ownership component, spatial assessments of changes in
population and number of households can be incorporated into the model. Finally, the
future of the southeast Texas landscape faces a number of uncertainties. As already
mentioned, continued sale of timberland will not only result in extensive loss of
forestland, but likely also disastrous effects on bear habitat. Several water diversion
projects are proposed that could disturb the balance of natural ecosystems. Also, a
proposed eight-lane Trans-Texas Corridor superhighway to connect Chicago and
Brownsville would bisect BTNP and nearby forestlands (http://www.dot.state.tx.us).
Possibilities of large-scale land use changes are making bear management planning more
difficult. Although sensitivity adjustments to the model did not result in significant
changes to the rankings, this does not guarantee that changes in feasibility rankings will
not occur in the future. Efforts to concentrate bear recovery efforts in areas that are least

likely to experience large-scale change will be critical for a successful reintroduction.

199

 

0:0: >080 80::
58:00 :000 80:: 0::0::0:8: :0 88:0:

::::00 :000

:0: 880:8: 8:86 80:: 0::0> 0w0:0><
::::00 :000

:0: 880:8: :0:/.86 80:: 020> 0w0:0><
::::00 :000

:0: 880:8: >036 80:: 020> 0:803“:

SOON Ba .82 .23:
0:0: 80:3: 86:00 m3: 88:60:

0:00: ::030: 8:880 03:60:

0>0: 0: :0:: 0:08 0:03 ::3 0::
8 :00: 060:: 0 :00: 0>0: 0:3 3:080:60:

6:00: ::030: 8:380 03:60:

0>0: 0: 30:: 0:08 0:03 0:00: 80:0
0308:2305: 0:08 0:03 0:3 3:080:60:

0:00: ::030: 8:880 03:60: 0>0: 0:
30:: 0:08 0:03 00:0 0:: 0: 38:60: :030 Z

0:00: ::030: 8:880 03:60: 0>0: 0:
30:: 0:08 0:03 9:080:80: 80:30:: 0:02

6::0::0:8: 8:8 :0::

800: ::030: 8:880 02:60: 0>0:

:23 8 :00: 00m

6:00: 80:0 03230::

88:60: :000: :0 8:85
8:23:00
0>_:0:00:::0-0>68:

 

::::00 :0 0:0 :0::08 :0 0w:0:0 :0 0:0: 0: 300:: 0:08 0:03 0::0::0:8: :0w::0> 0w<
0:0: 30085 86:00 me
ooom ::0 om:— 0w0:0>0 8:080: :0:: 6:00: ::030: 8::::::0
”0:08 0:03 0:3 38:60: ::::00 ::00:0: 03:60: 0>0: 0: :0:: 0:08 0:03 8:02 8:80
:0::0N::0:080:0: :0:08 :0: 00806 0:0Q 8:80:06: 0::0::0>

 

80:0:800 :0600 £088 :00:

0:: :0 8:086:00 :0: :06: 08.3 :0:: 6:00: 0:003 ::030: 8:880 88:00:: :0: 83060.: 8.86 ::00::::w6 ::00:6::0:m 6:. 2:0 :1

200

Table 7.2. Actual median age (from census data), verified median age (from simulation
of present) and estimated (simulated) changes in median age of residents within counties
of the study area for years 2010 and 2020.

 

 

 

Median age
County Actual Verified Year 2010 Year 2020
Angelina 34.2 34.4 37.6 41.0
Hardin 36.0 36.3 40.3 44.6
Jasper 37.3 37.6 41.3 45.4
Liberty 34.0 34.1 36.5 39.1
Newton 36.9 37.3 41.5 46.3
Orange 36.1 36.7 41.9 47.9
Polk 39.3 39.6 42.0 44.5
Sabine 47.0 47.8 52.7 58.1
San Augustine 42.1 42.3 45.2 48.3
San Jacinto 40.0 40.4 44.6 49.3
Trinity 43 .3 43.5 46.2 49.1
Tyler 38.9 39.2 41.0 42.8

 

201

Table 7.3. Actual percent males (from census data), verified percent males (from
simulation to present) and estimated (simulated) changes in percent males of resident

populations within counties of the study area for years 2010 and 2020.

 

Percent (%) of population that is male

 

 

County Actual Verified Year 2010 Year 2020
Angelina 48.2 49.1 50.0 52.8
Hardin 48.4 47.0 48.1 48.3
Jasper 47.3 48.2 47.6 47.0
Liberty 47.9 46.4 44.4 40.6
Newton 52.1 56.2 59.2 66.3
Orange 48.2 49.5 48.1 48.5
Polk 54.4 60.7 65.3 76.9
Sabine 46.7 48.1 47.1 47.2
San Augustine 46.1 47.5 46.9 46.1
San Jacinto 50.3 50.9 51.2 51.0
Trinity 48.2 48.2 48.7 64.5
Tyler 53.5 59.0 47.6 74.1

 

202

 

N N o N : : o N v : N N
N0: ::0 8: ..: 3: No 8: .: 00: No No .: 858
N N : N v o v N
a: No :3 ..N 3:: No a ..N :0: NN 3: ..N
N N : 0 N: N N N NN NN 0: N ::0:
SN 8: 8: ...N: ::::N 8: 8: ..N: SN 8: 8: ...N: 00::0
:. o N ::
N N : e: No: 8: N0 3
No: NN No a v o v N 8N 8: 8: ...o: 8:302
No: No 3 ...: No: 8: No ...:
N N : ::: v o N 0: SN 8: 8: ...: : 0:03:
No: 00: N0 .N
::::N 8: 8: ..N ::::N 8: 8: ...N N o N N 080:.
N : N v
8: 8: 3 .... :0: NN NN L. No: N0 N0 ..N
: o : N NN N: :N N N : : N :60:
N N :0 N
N N : N NN: :0 N0 ...0
a : N N NN N: NN N
::::N 8: 8: ..N 8: g N: ..N :N N: 0 v 2:009...
:06: N00 :00 V23: :08: N am ::0N :50 :8: Nam ::0m :53:
ONON SON some: 0550

 

:

 

06380 0603:9338 ::0 0:03:2386
cc: :0 N:0N 03: :0 :08: .N:0>000: :00: :0: 838300 :0 8:680: 6082000 Smom ::0 Bow 800.3 :0:00.N0:: ::0 808:: :1. 030:.

203

0020386 :000: 0:: :0: x00: 05 $050200 @520: :0:0>0 0:: 00:00:05 AL 02650 :5

 

00:

we

no:

00:

no

ma

mo

00:

no

oo—

00

oo

:
*0—

*N

*0

00:

00:

Va:

N—

\0

mm
5:

mo

mo

mm

mo

2:

No

Q:

00-

on

*O—

*N

*0

Vln

MVWOO

0m

m9

2

02

mm
o:

3
on

v—nhfi'v—t

00

2
mm
mm

ON
3
00

00

*5

*

WCNVmo—‘NMVWO

*v

('7

SSH

xzcth

0:503 00m

0:::Nsw:< 00m

38:53 .08 0:00P

204

Table 7.5. Present and projected (years 201 O and 2020) social rankings of counties for

bear recovery.

 

 

County Present 2010 2020
Angelina 7 7 5
Hardin 9 9 10
Jasper 2 1 1
Liberty 5 6 7
Newton 1 2 2
Orange 10 1 1 1 1
Polk 1 1 10 8
Sabine 6 5 6
San Augustine 8 8 9
San Jacinto 4 4 4
Trinity 12 12 12
Tyler 3 2 3

 

205

     

h

San Augustine

Figure 7 . 1. The 12 counties (outline, labeled) in southeastern Texas included within the
study area, as well as BTNP (solid shaded) and proclamation boundaries of four National
Forests (hatched).

206

Location

1

 

 

—' Timber management

 

 

I

Quaptity

Harvest strategy

 

 

 

Lb Bear habitat

— Quality

 

 

1
Distribution

 

V

Socioeconomics
(including attitudes)
|

—~ Social

1
Demography

 

 

 

 

 

V

@i-criteria analys)

I

 

Ranking
of
counties

 

 

 

Figure 7.2. Conceptual framework for ranking counties within the study area as
ecologically and socially feasible for a black bear reintroduction. Solid arrows indicate
the flow of data through the framework.

207

16

14

12

y.—
0

Frequency
00

 

 

 

 

 

 

 

 

 

 

-0.07 -0.05 -0.03 -0.01 0.01 0.03 0.07

Rate of change in timberland area (ha/year)

Figure 7.3. Histogram illustrating normally distributed function for rates of timberland
gains and loses across counties of the study area.

208

 

  

Angelina
timberland
current (I)

1

Angelina

change in
timberland
between t
and t + 1

J,

Angelina
—-h timberland
t + 1

Angelina rate of
timberland gain
or loss

   

 

   

 

 

 
 

 

 

 

 

 

 

 

 

 

 

Figure 7.4. Example model structure of the timberland component. “Angelina timber
current” indicated the area of timberland within Angelina County at time t. “Angelina
rate of timberland gain or loss” acted as criteria to designate whether timberland was
gained or lost at each time interval as determined by random number generation.
“Angelina change in timberland area between t and t + 1” quantified the total amount of
timberland gained or lost in Angelina County for each time interval. The total amount of
timberland gained or lost in time t was then added to “Angelina timberland current” to
calculate “Angelina timberland t + 1.” This model structure was repeated for each
county within the study area.

209

 

 

 

Angelina

 

 

 

 

  

Angelina rate of

    
   
 

 

 

 

 

 

 

 

 

timberland timberland gain
current or loss
Angelina
change in
timberland
area Angelina
betweent N-—-> habitat
and t + 1 current
Angelina An 1'
timberland h ge ma
t + 1 c ange in
habitat area
between t
and t + 1
Angelina ‘
habitat
t + 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 7.5. Example model structure of the habitat component. “Angelina habitat
current” indicated the area of total habitat within Angelina County at time t. “Angelina
change in habitat area between t + I” quantified the total amount of habitat gained or lost
in Angelina County for each time interval. The total amount of habitat gained or lost in
time t was then added to “Angelina habitat current” to calculate “Angelina habitat t + I .”
This model structure was repeated for each county within the study area.

210

 

Angelina
social
current

i

Angelina
changein
social
variable
between t
and t + I

J,

Angelina
—-—> social
1 + I

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 7.6. Example model structure of the social component. “Angelina social current”
indicated the current value of age and gender social variables used in the model at time t.
“Angelina change in social variable” acted as criteria to designate change in the social
variable at each time interval as determined by random number generation. “Angelina
change in social variable area for t and t + 1” quantified the total change in each social
variable in Angelina County for each time interval. The total amount of change in each
social variable in time t was then added to “Angelina social current” to calculate

“Angelina social t + 1.” This model structure was repeated for each county within the
study area.

211

00:0 >25 0::: mo 00.00500 :0 08:00 00:0 8010083 .00 00:08:00 Amway 008605 m:m:0> 0030000 .N..N. 0:::me

 

 

fiasco
L
0
990
0
cm
0:. n:
m
0..
00 m
cm W
m
00: W
0
cm: m
U1
0000605. 0: W
0020000-
00:
00:

 

 

212

00:0 030 0:0 00 0000300 :0 08:00 00:0 00:30:: 030000 >23: .00 00008000 Ammi 02200:: 000:?» 003030 .ws 053“:

   

x2500
@00 0&090 4V «V
.../Q 00,0 00,00 9% .000. .00 0.00 0%.. 0.02 .900 0%. 9%

0

ON

ov

o0

On

83:02,: I , 8:
005030 I om:

 

 

 

o:

(w 0001) 29m mtqeu alqmins KlqfitH

213

00:0 0030 0:0: .00 0000500 :0 000000 00:0 00050: 0300000 .00 0000000000 Ammi 00000005 0:0:0> 0030000 .06 Rama

50:00

 

, 83:02.: I .
03530 I, ON

(211 0001) we unmet! alqmins

 

 

 

mm

214

 

00:0 >030 05 .00 00000000 :0 000:00 00:0 0005000 000%:08 .00 000080000 Ammi 00000020 0:0:0> 00>:0000 .0 0 .0. 0::w0n0

 

 

 

 

5:000
900
0
0w .0
V. 007 @0 9nd #0 94? I?
00 r? o o o . 0v 9 0 Q
0.0..
. 0000000.:— I L W

 

”0020000 I _

 

1 ' 1"],- “ma.

 

..
. 10.00 60;;

 

0—

2

cm

mm

on

 

mm

 

(mi 0001) we 1811mm lemfimw

215

 

00:0 030 0:0 .00 00005000 :0 000:00 00:0 0003000 050000050 .00 000080000 Ammi 00000003 0:0:0> 0030000 .0 0.0 053.0

3:000

OW
'—

 

Oln
Nv—a

WC
MM

800020 I. .,
8:800 I . .0

O
<-

 

 

:3
(mi 0001) 89m 191!qu alqmmsun

IE
3

 

 

O
In

216

.omom 00:0 008 ::0 00:0 00:0 05 .00 000550 000:00 080 0000:0080 Amm 0 50:5 00:00.080 000 8008 0:000:00 .38 0:030

3:000

 

   

 

 

(mi 0001) 99m puellaqwtl

,ONONDH
EON!
..oooN!

 

 

om—

 

 

03

 

217

.88 0:0 0 SN :0 00:0 >030 05 00 0000:0000 000:00 00:0 00050: 0300000 3:300 Em 0” :0080 0000035 0:0 883 0:000:00 .2 .0 0::w0n0

 

 

 

 

 

5:000
60
V. 60900 900500. W 2 .9 A,
0
«00. 00.00 0.00 00% 00.0.00 .0.
00/f 64. 020/. a9 9%? 0% .va. 001 nova 00? 0% 002/.
, o
anon
0N m
m.
m
::N m
-- m
S. m
0
ON m
5820. 8 m
”::::NI, m
W88! ON

 

 

 

ow

218

.omom 0:0 0 0 om :0 00:0 00:00 0:00 .00 0000:0000 000:00 00:0 00:00: 0300000 Amm H :0080 00000.02: 0:0 800$ 0:000:00 +0.0 0:0w0n0

 

 

 

 

 

3:000

0.0

09

0., .0
09 o 0»
f A0 no 0 9¢ \( f »N« 00
0 o. 0... 00 on. 0 w 0. 0. ... o
«o/f 0.2% 010 90 a»? Avov «0%. 09 five? 00v 0000 90¢
- , o
m
00
2
on
ONONU mN

05ml

SONI 0N
mm

 

 

(mi 0001) we mzqeq alqmtns

219

.omom 0:0 Bow :0 00:0 0000 0:0 .00 00000000 000:00 00:0 00030: 00:0w:0:: Amm H 0025 0080.800 0:0 800$ 000005 .m 0 .N. 03w?—

5:000

 

 

 

 

 

 

 

 

ENoN—UH
EON-0 om
WOOON I.

(an 0001) 9919 191mm! lenifiww

 

 

 

mm

220

.cmom 0:0 0 SN :0 00:0 50000 0:00 00 00000000 000:00 00:0 00:50: 030000000 GE 0” 0:00:00 00:00.8:0 0:0 885 0:000:00 .0 _ .0 0:03.00

 

 

 

5:000
00
K0900 000.0%. ¢ b
0%.! 60.0. 00.02. 900.. 9029 0000 000 001 600 09% 0200 920%

o
N m
m.
m
::: w
m.
m...
m
2 m
HONoN n: m
MOMON .W cm m

:88 I:

 

 

 

mm

221

000050000 :0 000000 00 5:000 0000 :0.0 0:0—000:00 :0::? 00050: 0300000 5:3: .00 00:< ...3003 0 003
m 0-0 0:0 2.800005: 0 003 w-m 2.030010 00:0: 0 :0 0000000: 500300000 00000 :0 0000w00000 0:: 0:00 :0 0.0 .00 0:000 < 0000000050:
:000 00003 0 00 52500000 000000 0:0 0000w00000 :00 0w:00_:0: 5:000 :0 00:0:00000 0:0 $050500 :00 000000: 800 00:00 .N. 0 .n 000?:

 

 

 

 

 

 

 

 

. 580:0.
N Dwgh N O Aofiw WVV . .
::oNS 0 :NaN E V:: 0 8%.va 5000 33
S
m.
08.20 2:00:03. :00 m.
. a . 10
aNN 0N0 000:: :NNV 0N0 05:005. 000.0: 2:500 8382 m.
0..
m.
:N0N.:00 8:302 :NNN.::.0 0:00 0:00:
0:008 0:00: ::N0.N:0 90:00:. :00
3o: 05:82 00::

0:00:08: 00:00:80

222

SUMMARY

Many researchers separately studied ecological and social factors that may affect
potential wildlife reintroductions, but few have considered such factors simultaneously.
As human activity continues to penetrate previously rural areas, wildlife management
must focus on ways to meet the needs of wildlife while also considering socioeconomics.
For this dissertation, my research goal was to determine the feasibility of a Louisiana
black bear reintroduction in and around Big Thicket National Preserve (BTNP), Texas.
Because of BTNP’s small size and disjunct distribution, the preserve alone is not large
enough to fully meet the needs of a viable bear population. Therefore, quantity and
quality of bear habitat among nearby private timberlands and other public lands will be as
important for a bear population recovery as habitat within the preserve. I explored six
focal subjects: (1) ecological feasibility of a reintroduction, as determined by remotely
sensed imagery (Chapter 2), (2) assessment of the importance of private timberlands for
maintaining bear habitat over time (Chapter 3), (3) residents’ attitudes toward bears and
an increase in the black bear population size (Chapter 4), (4) preferred management
strategy for a black bear recovery (Chapter 5), (5) spatial distribution of individual
attitudes toward a black bear recovery (Chapter 6), and (6) development of a conceptual
framework to integrate ecological and socioeconomic factors that may affect a
reintroduction both at present and in the future.

Although previous studies have focused on bear habitat characteristics in
particular locations within southeast Texas, no studies have adapted a landscape

perspective to assess the area and distribution of bear habitat across the region. With an

223

estimated 1.3 million ha of highly suitable habitat, southeastern Texas contains enough
habitat for a viable bear population. Large areas of highly suitable habitat are well
connected, as suggested by negative correlations between habitat patch size and mean
nearest neighbor distance between patches of highly suitable habitat. Habitat data were
analyzed at both the 30m (raw scale of data) and l-ha (more appropriate scale for
studying bear ecology), and the most significant finding related to this was that remote
areas of habitat (>1 km from human activity), located mostly within the eastern portion of
the study area, were only visible using a l-ha scale. This research suggests that there is
enough habitat for a population of approximately 2,256 bears, but additional more-
detailed land cover analysis within particular locations will be needed to improve
accuracy of habitat analysis.

A majority of public land was classified as highly suitable. However, because of
the relatively small area of public lands, the widespread presence of timberlands across
the landscape will likely be an important factor in determining the maintenance of bear
habitat over time. For the past 30 years, timberland area has been decreasing over time.
This has important implications for bear recovery for the reasons that (1) the proportion
of timberlands containing highly suitable habitat has increased over time as a result of
increased attention to stand management for wildlife, and (2) >30% of land that has been
consistently highly suitable for bears over the past 30 years was contained among
timberlands. Louisiana-Pacific Corporation’s (LP) divestment in 2002 will likely affect
large areas of habitat between BTNP units, between BTNP and local national forests, and
along the Texas-Louisiana border. Although no detailed data are yet available for

assessing how LP’s divestment will affect bear habitat, land is being sold to the highest

224

bidder, which does not bide well for maintaining bear habitat in an area that is attractive
for suburban development and retirement communities.

As important as ecological feasibility of a reintroduction is social feasibility.
Because southeastern Texas contains a large human population, it is essential to
incorporate residents’ attitudes toward and opinions about a bear reintroduction into bear
planning and management. Considerable research has focused on socioeconomic and
demographic factors that affect attitudes toward wildlife and conservation. A survey was
completed here to determine whether particular socioeconomic factors affect residents’
attitudes toward black bears and preferred management strategies for bear recovery.
Males, younger respondents, respondents who were more knowledgeable about bears,
and those who participate more frequently in passive-appreciative activities related to
wildlife were more likely to have positive attitudes toward bears and support increasing
the bear population. Concerns about problems that bears may cause (e. g., pilfering in
garbage, perception that bears are a danger to people) was a significant predictor of
whether a respondent preferred a passive (non-human assisted) or active (human-assisted)
recovery strategy. Because a majority of respondents were male, and the survey sample
was disproportionably white/Caucasian compared to actual local demographics, further
collection of social data is necessary to assess attitudes of underrepresented groups, as
well as attitudes of particular user groups such as beekeepers and hunters.

Spatial analysis of respondents’ preferred management strategy resulted in two
significant clusters. One cluster of respondents in proximity to Angelina National Forest
(in San Augustine and Jasper Counties) indicated that they did not support a non-assisted

bear recovery strategy. Specific reasons cited by these respondents included that they

225

(the respondents) thought that a non-assisted recovery would take too long, and that they
wanted bears in the area now. Another cluster of respondents in Orange County
indicated that they did not support a recovery strategy that sought to exclude bears from
the area. Orange County is the most urbanized county within the study area, and even
though respondents indicated that they did not want to exclude bears from the area,
approximately 40% indicated that they were not sure exactly what management strategy
they preferred. Because a strong spatial distribution of attitudes was not identified,
public outreach will likely to be more of a challenge than if particular attitudes were
neatly clustered across the landscape.

To integrate ecological and social data, a framework for a systems model was
developed and used to assess ecological and social feasibility of a bear reintroduction
across the study area. Based on results from this research, past trend of declining
timberland area is likely to continue and may dictate overall changes in bear habitat over
time. If large areas of timberland are lost, there is a good possibility that such areas will
be replaced by residential, commercial, and industrial development. Ideally, the greatest
ecological and social feasibility would both be found within the same counties, but this
was not the case. Results from the systems model suggested that the greatest ecological
feasibility for a reintroduction existed in Sabine, Trinity, and San Augustine Counties,
but social feasibility was greatest in Jasper, Tyler, and Newton Counties. As a result,
national forests, which also contain expanses of core habitat, may be better potential bear
release sites than BTNP. Knowledge about black bears was the social variable for which
changes in the variable’s value most affected changes in model output. Therefore, efforts

to increase residents’ knowledge about bears may result in increased social feasibility

226

across the study area. The Texas Parks and Wildlife Department may use this
information to customize public outreach and management based on the ecological and
social feasibility within a particular county.

This dissertation is only the beginning of research aimed to determine whether
bears can be reintroduced into southeast Texas. Because of the coarse-scale analysis used
for both ecological and social components, finer detailed analysis is needed to verify
habitat data, as well as further explore residents’ attitudes and opinions about a bear
reintroduction. The East Texas black bear management plan, which was released in
2005, suggests public outreach as the main focus for bear recovery efforts over the next
15 years. Major changes are proposed for the southeast Texas landscape that could be
detrimental to bear recovery efforts. But by starting public outreach now, the Texas
Parks and Wildlife Department may be able to create enough interest and support for bear
recovery among citizen groups that bear management goals will be included in planning
for local development of human infrastructure. Although it is unknown at this time
whether and when bears will ever be released into southeast Texas, I have developed a
systems model to integrate ecological and social data for evaluating ecological and
socioeconomic feasibility of a species reintroduction, which may also be applied to other
wildlife conservation-related endeavors.

Data collected from this dissertation are not exhausted, and I offer three ideas of
future research. First, I will continue development of the systems model to incorporate
changes in land use in areas that are not publicly owned or managed by private timber
companies. Although expansion of human activity will be greatly affected by presence of

timberlands, exploration of land use changes in other locations within the study area may

227

lend insight into how land cover has changed in areas that were occupied by timberlands
in the past. Second, although detailed analysis of social variables that affect attitudes
toward bears has been completed, I will seek to apply risk perception analysis to survey
data. Several models for assessing acceptance capacity for wildlife have been developed,
and it would be interesting to compare how survey data related to bear recovery can (or
cannot) be incorporated into landscape analysis. Third, I will seek to explore how data
and results from this research can be expanded into greater detail to contribute to the
development of theory related to interdisciplinary research, improve policy and decision-
making for wildlife management, and our understanding of the dynamics of human-

dominated ecosystems.

228

APPENDIX

229

A. 1. SURVEY PRENOTICE

The prenotice was printed on Michigan State University Department of Fisheries and
Wildlife letterhead

26 January 2004

«fname» «lname»
«address I »
«city_name». «state_prov_code» «zip_postal_code»

In a few days you will receive a request in the mail to fill out a questionnaire for an important research
project being conducted by Michigan State University in cooperation with the Texas Parks and Wildlife
Department.

I am writing in advance to encourage your participation in this questionnaire, and because I understand that
many people like to know ahead of time that they will be contacted. This questionnaire is one part of a
study to learn about opinions of East Texas residents regarding black bear sightings in your area. This is a
very important study that will help researchers and natural resource managers manage bears with East
Texas residents’ diverse opinions and interests in mind.

If you have any questions about this project now or after you receive your questionnaire, please feel free to
call me toll free at 1-800-814-9289. If you have questions or concerns regarding your rights as a study
participant, or are dissatisfied at any time with any aspect of this study, you may contact Peter Vasilenko,
Chair of the University Committee on Research Involving Human Subjects (UCRIHS), 202 Olds Hall,
Michigan State University, by phone: (517) 355-2180, fax: (517) 432-4503, e—mail: ucrihs@msu.edu. or
regular mail: 202 Olds Hall, MSU, East Lansing, MI 48824-1046.

 

It’s only with the generous help of people like you that our research can be successful. Thank you in
advance for your time and consideration.

Sincerely,

Anita Morzillo
Project Manager

P.S. As our way of saying thanks for your participation, we will be enclosing a small gifi with your survey.
Remember, your survey will arrive in a couple of days. We look forward to hearing from you!

230

A. 2. SURVEY MAILING #1 COVERLETTER

The coverletter was printed on Michigan State University Department of Fisheries and
Wildlife letterhead

29 January 2004

«name»
«address»
«city», «state» «zip»-«four»

Recently. you should have received a notice about an upcoming survey about black bears in East Texas. I
am now writing to ask for your help in a study of East Texas residents being conducted by Michigan State
University in cooperation with the Texas Parks and Wildlife Department. This questionnaire is one part of
a study to learn the opinions and desires of East Texas residents about black bear sightings in East Texas.
This very important study will help researchers and natural resource managers manage bears with East
Texas residents’ diverse opinions and interests in mind.

We are contacting a random sample of East Texas residents to ask their opinions about wildlife, black
bears, and black bear management. We are interested in the wide range of opinions that currently exist.
There are no right or wrong answers. m opinions are very important to us and v_v_ifl make a difference for
bear management in East Texas.

Your responses will be kept completely confidential. The survey has identifying information for mailing
purposes only. Your name and address will never be associated with your responses and your privacy will
be protected to the maximum extent allowable by law. Your response to this survey and any of the
questions is completely voluntary and it should take about 15 minutes to complete. You indicate your
voluntary agreement to participate by completing and returning this survey. Please complete this
questionnaire at your earliest convenience, seal it, and drop it in any mailbox (no envelope is needed).
Return postage has been provided.

If you have any questions about this project, please feel free to call me toll free at 1-800—814-9289. If you
have questions or concerns regarding your rights as a study participant, or are dissatisfied at any time with
any aspect of this study, you may contact Peter Vasilenko, Chair of the University Committee on Research
Involving Human Subjects (UCRIHS), 202 Olds Hall, Michigan State University, by phone: (517) 355-
2180, fax: (517)432-4503, e-mail: ucrihs@msu.edu. or regular mail: 202 Olds Hall, MSU, East Lansing,
MI 48824-1046.

Thank you very much for helping with this important study.

Sincerely,

Anita Morzillo
Project Manager

P.S. This survey is intended for someone who is at least 18 years of age and a resident of East Texas. If >
the person to whom this is addressed does not fit this description, please give this survey to a person in your
household who does. If no one in your household fits this description, please write on the survey that no
one was eligible to complete it and send the survey back to me. Many thanks.

231

A. 3. SURVEY INSTRUMENT fl

ATTITUDES ABOUT BLACK BEARS IN EAST TEXAS
1. Which of the following statements Lie—st describes your current interest in and
involvement with wildlife? (Please check [\l] ONE statement)

[ ] I am interested in wildlife AND I actively take part in wildlife related activities

[ ] I am interested in wildlife BUT I do no_t_ take part in many activities that are
specifically related to wildlife

[ ] I am NOT interested in wildlife AND I do mt take part in many activities that
are specifically related to wildlife

[ ] I am NOT interested in wildlife BUT for various reasons I am involved in
wildlife-related activities

2. Please indicate how often you, or members of your household, participate in each
of the following activities for work or pleasure. (Please circle ONE response for
ggc_h_ activity)

Ofien Sometimes Never

 

Hike

 

J og/run outside

 

Bike (trail/mountain or road)

 

Camp (tent, trailer, RV)

 

Motorboat/ j etski/waterski

 

Canoe/kayak

 

Ride motorized all-terrain vehicles

 

Read about wildlife

 

Watch wildlife TV shows or movies

 

Observe or study wildlife outdoors

 

Hunt big game (e.g., deer)

 

Hunt smaller animals (e. g., squirrel)

 

Fish

 

Work on a farm or ranch

 

Work in the timber industry

 

Work in the oil/ gas industry

 

 

 

Other activities (please specifiz)

 

 

 

 

232

3. Are you a member of any organizations related to wildlife (e.g., Ducks Unlimited,
Audubon Society)? (Please check N] ONE)

] Yes -- Please answer question 3a

I
[ ] No -- Please go to guestion 4

3a. If yes, what organization(s)?

 

 

 

4. Where do you get most of your information about wildlife in Texas? (Please
check [\I] ONE)

] Television
] Newspaper
] Magazines
] Internet
] Hunting/fishing regulation guide
] Wildlife professionals

] Family, friends, or co-workers

] Your own experiences

[
l
[
[
[
[
[
l
[ ] Other (please sped/52)

5. Please indicate which, if any, of the following types of interactions with black
bears you, or members of your household, have experienced in any location.
(Please check [\l] 2!! that apply)

Watched black bears in the wild or in captivity (e.g., in a zoo)
Read something or watched TV shows/movies about black bears
Hunted black bear

Had a personal encounter with a black bear

Livestock had an encounter with a black bear

Read or heard of a black bear being killed by authorities

None of the experiences described above

Other types of experiences (please specify below)

 

 

233

6. Have m ever seen a black bear in the wild? (Please check N] ONE)
[ ] Yes -- Please answer questions 6a and 6b
[ ] No -- Please go to question number 7

[ ] Unsure —- Please go to question number 7

6a. If yes, when and where?

 

 

6b. If yes, for each item below please indicate, on a scale of 1 to 5, your reaction
to seeing a black bear. (Please circle ONE number per line)

Happy 5 4 3 2 1 Unhappy
Excited 5 4 3 2 1 Not excited
Curious 5 4 3 2 1 Not curious
Frightened 5 4 3 2 1 Not frightened

7. Prior to receiving this survey, were you aware that: (Please circle ONE response
for each statement)
Yes No

 

Until the early 1900’s, East Texas contained a large
population of black bears

 

The number of black bear sightings in East Texas has
increased during the past decade

 

Black bear populations are increasing in size in
Arkansas, Louisiana, and Oklahoma

 

Black bears in East Texas are protected by both federal
and state legislation

Black bears exist throughout most of the United States
and North America

Black bears are mainly vegetarians

 

 

 

 

 

 

 

Please go on to the next page

234

8. Which of the following statements pg! reflects how you feel about black bears in
East Texas? (Please check N] ONE)

[ ] I would enjoy having black bears around AND I would my worry about problems
they may cause

[ ] I would enjoy having black bears around BUT I would worry about problems
they may cause

[ ] I would no; enjoy having black bears around BUT I would n_o_t worry about
problems that they may cause

[ ] I would n_ot enjoy having black bears around AND I would worry about problems
they may cause

[ ] I have no particular feelings about black bears regardless of problems caused or
not caused by them

9. It is likely that people in East Texas have many different opinions about black
bears. To what extent do you agree or disagree with each of the following

statements? (Please circle ONE response for each statement)
Strongly Strongly
Agree Ange Unsure Disagree Disagree

 

The presence of black bears is a sign
of a healthy environment

 

Black bears would reduce the size of
wild hog populations

 

Black bears in East Texas would
increase my quality of life

 

Black bears near my home would
increase my quality of life

 

Black bears have the right to exist
wherever they may occur

 

I would feel personally at risk if black
bears exist in East Texas

 

I am afraid of black bears

 

Black bears commonly harm humans

 

Wildlife experts know how to manage
black bears

Wildlife experts understand
landowners’ concerns about black
bears

 

 

 

 

 

 

 

 

 

235

 

10. To what extent do you agree or disagree with each of the following statements?
(Please circle ONE response for each statement).

Strongly Strongly
Agree Agree Unsure Disagree Disagree

 

The black bear population in East Texas
should be increased

 

The black bear population in East Texas
should be increased only if steps are taken to
lessen the chances of human-bear conflicts

 

The black bear population in East Texas
should remain the same

 

Black bears should mt exist in East Texas

 

 

 

 

 

 

 

11. Do you hunt? (Please check [\/I ONE)
[ ] Yes -- Please answer question 11a

[ ] No -- Please go to question 12

11a. Would you be interested in hunting for black bear in East Texas?
(Please check [x11 ONE)

[]Yes
[]NO

[ ] Unsure

12. In general, do you believe black bears are a potential danger to humans? (Please
check N] ONE a_m_l answer 12a)

[]Yes
[]NO

[ ] Unsure

12a. Provide reasons for your answer. (Please be as specific as possible)

 

 

236

13. Do you believe black bears are a nuisance? (Please check [‘1] ONE M answer

13a)
[]Yes
[]No

[ ] Unsure

13a. Provide reasons for your answer (Please be as specific as possible)

 

 

14.

Do you think black bear populations in East Texas should increase naturally
(i.e., without assistance from a natural resource agency)?
(Please check N] ONE M answer 14a)

[]Yes
[]No

[ ] Unsure

14a. Provide reasons for your answer (Please be as specific as possible)

 

 

15.

Do you think that natural resource agencies should assist in increasing the black
bear population size in East Texas? (Please check N] ONE any! answer 15a)

[]Yes
[]No

[ ] Unsure

15a. Provide reasons for your answer (Please be as specific as possible)

 

 

16.

Would you support the restocking of black bears into suitable habitats in East
Texas by natural resource agencies? (Please check [‘1] ONE)

[]Yes
[]No

[ ] Unsure

237

17. Some people are concerned that recovery of black bears in East Texas may cause
problems for people. In your opinion, which statement best describes how potential
bear problems should be handled by natural resource agencies? (Please check [\l]
ONE)

] Relocate the problem bear to a different area

] Use dogs to frighten bears until bears are no longer a problem

] Kill problem bears only after several offences
] Bears should not be disturbed

[

l

[ ] Kill problem bears after a first offense
I

[

[ ] Other (please specify)

 

 

18. Is each of the following more or less like] to ha en to on compared to you
having a negative encounter with a black bear? For example, if you believe that
you are more likely to become a murder victim than experience a negative
encounter with a bear, then circle more likely. (Please circle ONE response for
gay! item)

 

 

 

. - I cannot determine this
Example. Becoming a murder More Less Unsure based on my personal
vzctlm. Likely Likely

experience

 

Getting into a car accident

 

Getting into an accident if
you ride a motorcycle

 

Getting scratched by a cat

 

Having an accident if you
drive a tractor

 

Getting into a commercial
airline crash

 

Dying as a result of cancer

 

Getting bitten by a dog

 

Getting struck by lightning

 

Winning $1 million in the
lottery

 

 

 

 

 

 

 

238

In order for us to better understand peoples’ responses to the previous questions, we
need to know a few things about your background. Please remember, your
responses are completely confidential. Neither your name nor your address will be
directly linked to your responses in any way.

19. For approximately how many years have m:

Lived in East Texas? years

Lived at your current residence? years

20. For approximately how many generations has your family lived in East Texas?

generations

 

21. In what county do you live? County

22. If you have lived in other locations, please indicate the states or
countries that you have lived in (including other regions of Texas).

 

 

23. Do you own any land in East Texas? (Please check [\l] ONE)

[ ] Yes -- Please answer questions 23a and 23b

[ ] No -- Please go to question 24

23a. For how many years have you owned land in East Texas?

years

 

23b. How many acres of land do you own? acres

24. Which of the following land use activities presently take place on the land that
you own g live on? (Please check [\l] gl_l that apply)

[ ] Timber management [ ] Row crop agriculture (e.g., corn)
[ ] Livestock grazing [ ] Residential

[ ] Hunting [ ] Oil/natural gas extraction

[ ] Beekeeping [ ] All-terrain vehicle use

[ ]Commercial/industrial
[ ] Other (please specify)

 

 

239

25. Which of the following land use activities do you expect will take place in the
future on the land that you own 21; live on. (Please check [\l] ng that apply)

[ ] Timber management [ ] Row crop agriculture (e. g., corn)
[ ] Livestock grazing [ ] Residential

[ ] Hunting [ ] Oil/natural gas extraction

[ ] Beekeeping [ ] All-terrain vehicle use

[ ] Commercial/industrial

[ ] Other (please specify)

 

 

26. In what type of community do you currently live? (Please check [\l]
ONE)

[ ] Rural, Farm

[ ] Rural, Non-farm

[ ] Small town (<5,000 people)

[ ] Large town (5,000-10,000 people)
[ ] Suburb

[ ] Small city (10,001-50,000 people)
[ ] Large city (>50,000 people)

27. How many individuals are in your household?

How many individuals are less than 18 years old?

 

28. Do you have any pets? (Please check [‘1] ONE)
[ ] Yes
[ ] No

29. Are you male or female?
[ ] Male

[ ] Female

30. In what year were you born? 19

240

31. What is the highest level of formal education that you have
completed? (Please check N] ONE)

Primary school (grade 8)

High school graduate or equivalent (e. g., GED)
Vocational or trade school

Some college

Associate’s degree (2 year degree)

College graduate (Bachelor’s or 4 year degree)

I
l
[
l
l
l
I

HHHHHHH

Graduate or professional degree

32. What is your race or ethnicity? (Please check [‘1] all that apply)
[ ] White/Caucasian
[ ] Black or Afi'ican American
[ ] Hispanic or Latino

[ ] Asian

[ ] Native American

[ ] Native Hawaiian or other Pacific Islander

[ ] Other (please specify)

 

33. What was your gross household income (before taxes) in 2003? (Please check N]
ONE)

[ ] Less than $20,000
[ ] $20,000 to $39,999
[ ] $40,000 to $59,999
[ ] $60,000 to $74,999
[ ] $75,000 or more

Please go on to the next page

241

34. For our future reference, by which means, if any, would you like to receive
information about black bears in East Texas? (Please check [\l] all that apply)

[ ] Pamphlet/brochure [ ] I am not interested
[ ] Compact disk (CD)

[ ] VCR tape

[ ] Digital video disk (DVD)
[ ] Public information session
[ ] Internet

[ ] E-mail

[ ] Other (please specify)

 

 

Please use this space for any additional comments.

[ ] Check [\l] here ifyou would like us to send you a summary of survey results.

[ ] Check [\l] here if you would like us to send you general information about black
bears in East Texas.

TO RETURN THIS SURVEY, SIMPLY SEAL IT AND DROP IT
INTO ANY MAILBOX. RETURN POSTAGE IS PROVIDED.

THANK YOU FOR YOUR TIME AND ASSISTANCE!!

242

A. 4. REMINDER POSTCARD

12 February 2004 w

Recently you were mailed a questionnaire from Michigan State University, in
cooperation with the Texas Parks and Wildlife Department, seeking your opinion
about black bears.

 

If you have already completed and returned the survey, please accept our sincere
thanks! If not, please do so today. Because of the small number of people contacted,
it is very important that we receive your feedback.

If you did not receive the questionnaire, or it got misplaced, please call me toll free
at 1-800-814-9289 and we will get another questionnaire in the mail to you.

Sincerely,

Anita Morzillo
Project Manager

243

A. 5. SURVEY MAILING #2 COVERLETTER

The coverletter was printed on Michigan State University Department Of Fisheries and
Wildlife letterhead

26 February 2004

«name»
«address»
«city», «state» «zip»-«four»

A few weeks ago we sent you a questionnaire asking for your opinions about wildlife, black bears, and
black bear management. So far, we have not received your response. If this letter and your completed
survey have crossed in the mail, please disregard this letter and accept our sincere thanks for your
participation in this study!

The comments of people who have already responded show that East Texas residents hold a wide variety of
views about black bears. We know the results are going to be very useful to natural resource managers and
researchers.

We are writing again because of the importance that your questionnaire has for helping to get accurate
results. Although we sent questionnaires to people across several counties, it is only by hearing from
nearly everyone to whom surveys are mailed that we can be sure that the results are truly representative of
the opinions of East Texas residents. If you have not had a chance to fill out the survey questionnaire, we
would appreciate your prompt attention.

Some people have told us that they have no interest in, knowledge of, or experience with black bears and
they feel that their response is therefore not important. We want you to know that everyone's opinion is
important for this survey. A few people have written to say that they should not have received the
questionnaire because they are not a resident of East Texas or they are not at least 18 years of age. If either
of these concerns applies to you, please give the survey to a person in your household who meets these
criteria. If no one in your household is eligible, please indicate this on the survey and send it back to us.
We would really appreciate it, and this way we can remove you from our mailing list.

Your name will never be associated with your responses in any way and your privacy will be protected to
the maximum extent allowable by law. Your response to the survey and any of its questions is completely
voluntary. The survey should only take about 15 minutes to complete. After you are finished completing
the questionnaire, please seal it, and drop it in any mailbox (no envelope is needed). Return postage has
been provided.

If you have any questions about this project, please feel free to call me toll free at 1-800-814-9289. If you
have questions or concerns regarding your rights as a study participant, or are dissatisfied at any time with
any aspect of this study, you may contact Peter Vasilenko, Chair of the University Committee on Research
Involving Human Subjects (UCRIHS), 202 Olds Hall, Michigan State University, by phone: (517) 355-
2180, fax: (517)432-4503, e-mail: ucrihs@msu.edu, or regular mail: 202 Olds Hall, MSU, East Lansing,
MI 48824-1046.

Sincerely,

Anita Morzillo
Project Manager

244

:0w0002 808:0

2:082 005:

.50:00:_m

800:0 0::: ::3 00 008 0: 0::: 0:: w:_0_0: :00 00» 030:0. 080002000 5:00:w 0: :0::0:00000 :00>

0080000: :00» 0: 000::

0: :0>0: :05 080: :00> 5:0 0080:00 wE=0E :00 :0:::0: 800000000: :0 00: 0:00:000 0: 0. .5:0_::00_0:00 0:200:00 00 00:0000 0: 00:: 00>
80:80:: :_ 000:00 :00 0:0 03 0: :0::0:::00:_ 0::: :00 00» w:_0_00 0: :0: 0.003 03 :0: .5082?» 0: 0:00:000 0::: 0: 080000: :00» .0800: 0<
000:0 :00 :00 00008805 030003 00320 :05 :_ 000000: 05080 00 :000 00 00 0: 0_:: E30: 0: 0::: 0::

820: 00% 0:060:000 50:0050 0.003 0 3 0:00:80 0:: :00 ::0 0: 00:05:: m :0:: 0:00: 0: 0000: 005:0 :0 00:: 0:: :_ 00:0 0:0 .:00:00 .:00 ::0 0:

00% :00 0:00:000 00002000 0006000 0 00:00::0 0>0: 03 .:0::0.m 5000320 00% ::00 03 00300 0:: 00.: m:_::%:0 :00 ::0 0: 00» w::0_00 :0: 0:0 0 >9

.00:0000:-:0: 0::: :00 000000: 0:: 0:80:08: :00 0B 00 00000000 300 0 00% :00 0: 0:: 0.003 03 .0000: 00: 0B :0:: :030_ 003 00200
0::: 0: 0:0: 00:0000: :00 00:5 0000.0 :00m :_ ::0E0w0:0:: :00: 0000—: :00:0 :0_:_00 :00» 83000 0:_0::0_:0000 0 00.00:: 0:03 00» 5:00000

voom 0:00. 0m

>m>m3m 003.520.0409:— m—mZOmmmy—ZOZ .0 .<

245

.0 0

. O Jaws—H. 02 H 0 00> H 0 gma: 0:000: ::0: 00.» 0:03 .0

0.080“. H 0 0.00). H 0 0.0.0800 :0 0.0:: 00.» 0:< .w
02 H . 00> H 0 000:000 002.00.00.00: :0.» 0G .0

 

2:000 . . oz . . 00> . .
.000 «00:90 0.00.8000 00:88: 020:0: 5. 0000.0. :00H 0. 80:30: 0.08.00 0::. 0:00: 0.00.: 00 0:300:00: 0:: ::00000 000 0.003 .0

as0:0 . . oz . . 00> . 0
.000 0:00:90 00:08:: 0: :00000 .0.::0:00 0 0:0 0:00: 0.00.: 08:0: 00.» 00 ..0:0:0u :— .m

0:083 H H 07. H H 00> H 0
.000 0000.00 0005.: :0: a: 0:02: 0.00... 0:30.. 3.05 .3: 2.53 ...

0:083 H . 07. H 0 00 > H . 0:0.:0:0w0> 50.08 0:0 0:00: 0.02m. .0

0:00: 3 H . 02 H . 00 > H . 000000 :000 0:: w:.:00 00808:. 00: 0000... :00m. :. 0w:.::w.0 :00: 0.00.: 00 :0:::0: 0: ... .0

0:08: H . 07. H . 00 > H . 00.:0::< :::02 0:0 00:0:m 00:83 0:: 00 :88 :00:w00::: :008 0:00: 0.005 .:

0:083 H 0 02 H 0 00 > H 0 0:00: 0.00.: 00 8.0.0000 0W0: 0 08.088 00:80. :00m .0600. 5:00 0:: :::D .0
::::05030 a :0\080&.0: § 0000.00 08030 ":0:: 0:030 00.: 0:03 $0.080 0...: 0000.000: 0: 80:0 .m

3.2.0., 50:20.0 80:92: 020 . . 050: . 0 020500 . . 05303 0:28 0 0 0:20: . . 0:05.: . 0

30%: :0::: :00 0:00:90 0.005.500 00.30.30 0:: 00 0:0 :. 0:005:80 00» 00 £00.» .0305 0 :0 .N

 

”30.0000 0000.00 :0000: :0::0:< H .
08:00.0: 5: 00:00:80 8.0: :000:0 0.003 . H 0 00300 0:: 002000: :0>0: . H .

00500 0:: :. 00000. 0:: :. 00:00:08. :0: E0 . H 0 0:00: :00:0 0.0.. 50> 30:0. . H .

0E0:wo:0 0:. 80:0 ::0E::000Q 00:0..3 0% 00:00 00:00 0:: ::3 0000::0 ::0 . H . 000500 :0380 0: 0::: 0>0: :0: 00 . H .
00:00..0::00 :0\0:0 w:0. 00: 000.00. 00200 0: ... H 0 000200 w:.:0380 00... :0: 00 . H H

0.0000 :0::: :00 0:00:50

...00000. :00m 0. 0:00m 0.005 :00:0 0000:.::< :0 00:00 :00 0: 000000: :0: 0.0 :0» 5.3 00000: :0.00.::00 0 0.5:: 00>? .:

00.050: .0 .0.

246

LITERATURE CITED

Abernathy, F. E., editor. 1967. Tales from the Big Thicket. University of North Texas
Press, Denton, Texas, USA.

Allnutt, T. F ., W. W. Wettengel, J. V. Reyna, R. C. de Leon Garcia, E. E. Ifiigo Elias, and
D. M. Olson. 2002. The efficacy of TM satellite imagery for rapid assessment of

Chihuahuan xeric habitat intactness for ecoregion-scale conservation planning.
Journal of Arid Environments 52: 135-153.

Anderson, D. R. 1997. Corridor use, feeding ecology, and habitat relationships of black
bears in a fragmented landscape in Louisiana. Thesis, University of Tennessee,
Knoxville, Tennessee, USA.

Anderson, S. H., and B. J. Crompton. 2002. The effects of shelterwood logging on bird
community composition in the Black Hills, Wyoming. Forest Science 48: 365—
372.

An. L., J. Liu, Z. Ouyang, M. Linderman, S. Zhou, and H. Zhang. 2001. Simulating
demographic and socioeconomic processes on household level and implications
for giant panda habitats. Ecological Modelling 140: 31—49.

Arcese, P., and A. R. E. Sinclair. 1997. The role of protected areas as ecological
baselines. Journal of Wildlife Management 61: 587—602.

Ash, S. J ., and C. E. Adams. 2003. Public preferences for free-ranging domestic cat
(F elis catus) management options. Wildlife Society Bulletin 31: 334-339.

Babbie, E. 1990. Survey research methods. Second edition. Wadsworth Publishing
Company, Belmont, California, USA.

Bailey, S. A., R. H. Haines-Young, and C. Watkins. 2002. Species presence in
fragmented landscapes: modeling of species requirements at the national level.
Biological Conservation 108: 307-316.

Bailey, T. C., and A. C. Gatrell. 1995. Interactive spatial data analysis. Pearson
Education Limited, Harlow, Essex, England.

Bath, A. J. 1987. Attitudes of various interest groups in Wyoming toward wolf
reintroduction in Yellowstone National Park. Thesis, University of Wyoming,
Laramie, Wyoming, USA.

Bath, A. J. 1989. The public and wolf reintroduction in Yellowstone National Park.
Society and Natural Resources 2: 297—306.

247

Bath, A. J. and T. Buchanan. 1989. Attitudes of interest groups in Wyoming toward
wolf restoration in Yellowstone National Park. Wildlife Society Bulletin 17: 519—
525.

Beausoleil, R. A. 1999. Population and spatial ecology of the Louisiana black bear in a
fragmented bottomland hardwood forest. Thesis, The University of Tennessee,
Knoxville, Tennessee, USA.

Beecham, J. J. 1980. Population characteristics, denning, and growth patterns of black
bears in Idaho. Dissertation, University of Montana, Missoula, Montana, USA.

Beeman, L. E, and M. R. Pelton 1976. Homing of black bears in the Great Smoky
Mountains National Park. International Conference on Bear Research and
Management 3: 87—95.

Beier, P. 1993. Determining minimum habitat areas and habitat corridors for cougars.
Conservation Biology 7: 94-108.

Benson, J. F. 2005. Ecology and conservation of Louisiana black bears in the Tensas
River Basin and reintroduced populations. Thesis, Louisiana State University,
Baton Rouge, Louisiana, USA.

Black Bear Conservation Committee (BBCC). 1997. Black bear restoration plan. Baton
Rouge, Louisiana, USA.

Blanchard, B. M., and R. R. Knight. 1995. Biological consequences of relocating grizzly
bears in the Yellowstone ecosystem. Journal of Wildlife Management 59: 560—
565.

Boardman, A. E., A. R. Vining, D. L. Weimer, D. H. Greenberg. 1996. Cost-benefit
analysis. Prentice Hall, Upper Saddle River, New Jersey, USA.

Bolen, E. G., and W. L. Robinson. 1999. Wildlife ecology and management. Second
edition. Prentice Hall, Upper Saddle River, New Jersey, USA.

Bolger, D. T., A. C. Alberts, R. M. Sauvajot, P. Potenza, C. McCalvin, D. Tran, S.
Mazzoni, and M. Soulé. 1997. Response of rodents to habitat fragmentation in
southern California. Ecological Applications 7: 552-563.

Bowker, B., and T. Jacobson. 1995. Louisiana black bear (Ursus americanus luteolus)
recovery plan. US. Fish and Wildlife Service, Atlanta, Georgia, USA.

Bowman, J. L. 1999. An assessment of habitat suitability and human attitudes for black

bear restoration in Mississippi. Dissertation, Mississippi State University,
Mississippi State, Mississippi, USA.

248

Bowman, J. L., B. D. Leopold, F. J. Vilella, D. A. Gill, and H. A. Jacobson. 2001.
Attitudes of landowners toward American black bear compared between areas of
high and low bear populations. Ursus 12: 153—160.

Bowman, J. L., B. D. Leopold, F. J. Vilella, and D. A. Gill. 2004. A spatially explicit
model, derived from demographic variables, to predict attitudes toward black bear
restoration. Journal of Wildlife Management 68: 223—232.

Brashares, J. S., P. Arcese, and M. K. Sam. 2001. Human demography and reserve size
predict wildlife extinctions in West Africa. Proceedings of the Royal Society of
London Series B-Biological Sciences 268: 2473-2478.

Brody, A. J ., and J. N. Stone. 1985. Timber harvest and black bear population dynamics
in a southern Appalachian forest. International Conference on Bear Research and
Management 7: 243—250.

Brody, A. J ., and M. R. Pelton. 1989. Effects of roads on black bear movement in
western North Carolina. Wildlife Society Bulletin 17: 5—10.

Breitenmoser, U., C. Breitenmoser—Wursten, L. N. Carbyn, and S. M. Funk. 2001.
Pages 241—281 in J. Gittleman, S. M. Funk, D. Macdonald, and R. K. Wayne,
editors. Carnivore conservation. Cambridge University Press, Cambridge,
England.

Brown, G. 2005. Mapping spatial attributes in survey research for natural resource
management: methods and applications. Society and Natural Resources 18: 17—
39.

Brotons, L, M. Monkkonen, and J. L. Martin. 2003. Are fragments islands? Landscape

context and density-area relationships in boreal forest birds. The American
Naturalist 162: 343-357.

Brown, G. 2005. Mapping spatial attributes in survey research for natural resource
management: methods and applications. Society and Natural Resources 18: 17—

39.

Brown, J. H. 1971. Mammals on mountaintops: nonequilibrium insular biogeography.
The American Naturalist 105: 467—478.

Bruner, A. G., R. E. Gullison, R. E. Rice, and G. A. B. Fonseca. 2001. Effectiveness of
parks in protecting tropical biodiversity. Science 291: 125—128.

Buckland, S. T., and D. A. Elston. 1993. Empirical models for the spatial distribution of
wildlife. Journal of Applied Ecology 30: 478—495.

Burton, T., D. Koch, D. Updike, and A. Brody. 1994. Evaluation of the potential effects

249

of sport hunting on California black bears. International Conference on Bear
Research and Management 9: 231—235.

Cardillo, M. D. W. Macdonald, and S. P. Rushton. 1999. Predicting mammal species
richness and distributions: testing the effectiveness of satellite-derived land cover
data. Landscape Ecology 14: 423-425.

Carroll, 0, R. F. N083, and P. C. Paquet. 2001. Carnivores as focal species for
conservation planning in the Rocky Mountains region. Ecological Applications
11: 961-980.

Chapin, F. S., E. S. Zavaleta, V. T. Eviner, R. L. Naylor, P. M. Vitousek, H. L. Reynolds,
D. U. Hooper, S. Lavorel, O. E. Sala, S. E. Hobbie, M. C. Mack, and S. Diaz.
2000. Consequences of changing biodiversity. Nature 405: 234—242.

Chapin, T. G., D. J. Harrison, and D. D. Katnik. 1998. Influence of landscape pattern on

habitat use by American marten in an industrial forest. Conservation Biology 12:
1327-1337.

Clark, J. D. 1991. Ecology of two black bear populations in the Interior Highlands of
Arkansas. Dissertation, University of Arkansas, Fayetteville, Arkansas, USA.

Clark, J. D., D. Huber, and C. Servheen. 2002. Bear reintroductions: lessons and
challenges. Ursus 13: 335—345.

Clark, J .D., D.L. Clapp, K.G. Smith, and TB. Wigley. 1991. Black bear damage and
landowner attitudes toward bears in Arkansas. Proceeding of the Annual
Conference of Southeastern Association of Fish and Wildlife Agencies 41: 432-
437.

Clark, J. D., J. E. Dunn, and K. G. Smith. 1993. A multivariate model of female black
bear habitat use for a geographic information system. Journal of Wildlife
Management 57: 519-526.

Clark, J. D., D. Huber, and C. Servheen. 2003. Bear reintroductions: lessons and
challenges. Ursus 13: 335—345.

Coluccy, J. M., R. D. Drobney, D. A. Graber, S. L. Sheriff, and D. J. Witter. 2001.
Attitudes of central Missouri residents toward local giant Canada geese and
management alternatives. Wildlife Society Bulletin 29: 116-123.

Comly, L. M. 1993. Survival, reproduction, and movements of translocated nuisance
black bears in Virginia. Thesis, Virginia Polytechnic Institute and State
University, Blacksburg, Virginia, USA.

Cork, S. J ., I. D. Hume, and W. J. Foley. 2000. Improving habitat models and their

250

utility in koala conservation. Conservation Biology 14: 660—668.

Cortina, J. M. 1993. What is coefficient alpha? An examination of theory and
applications. Journal of Applied Psychology 78: 98-104.

Craven, S. R., D. J. Decker, W. F. Siemer, and S. E. Hygnstrom. 1992. Survey use and
landowner tolerance to wildlife damage management. Transactions of the North
American Wildlife & Natural Resources Conference 57: 75—88.

Cuaron, A. D. 2000. Effects of land-cover changes on mammals in a neotropical region:
a modeling approach. Conservation Biology 14: 1676—1692.

Decker, D. J ., and N. A. Connelly. 1989. Motivations for deer hunting: implications for
antlerless deer harvest as a management tool. Wildlife Society Bulletin 17: 455-
463.

Decker, D. J ., T. L. Brown, D. L. Hustin, S. H. Clarke, and J. O’Pezio. 1981. Public
attitudes toward black bears in the Catskills. New York Fish and Game Journal
28: 1—20.

Decker, D. J. and J. O'Pezio. 1989. Consideration of bear-people conflicts in black bear
management for the Catskill Region of New York: application of a
comprehensive management model. Pages 181-187 in Proceedings of a
Symposium on Management Strategies. Northwest Territories Department of
Renewable Resources, Canada.

Decker, D. J ., R. A. Smolka, Jr., J. O’Pezio, and T. L. Brown. 1983. Social determinants
of black bear management for the northern Catskill Mountains Pages 239—247 in
Beasom, S. L., and S. F. Roberson, editors. Game Harvest Management. A
proceedings of the Third International Symposium of the Caesar Kleberg Wildlife
Research Institute. College of Agriculture, Texas A&I University, Kingsville,
Texas. 3—6 October, 1983.

Diamond, J. M. 1975. The island dilemma: lessons of modern biogeographic studies for
the design of natural preserves. Biological Conservation 7: 129-146.

Diamond, J. M. 1976. Island biogeography and conservation: strategy and limitation.
Science 193: 1027-1029.

Dillman, D. 2000. Mail and Internet surveys: the tailored design method. Second
edition. John Wiley and Sons, Inc., New York, New York, USA.

Dobey, S., D. V. Masters, B. K. Scheick, J. D. Clark, M.R. Pelton, and M. E. Sunquist.

2005. Ecology of Florida black bears in the Okefenokee-Osceola Ecosystem.
Wildlife Monographs 158: 1-41.

251

Donovan, T. M., P. W. Jones, E. M. Annand, and F. R. Thompson 111. 1997. Variation
in local-scale edge effects: mechanisms and landscape context. Ecology 78:
2064-2075.

DuBrock, C. W., A. R. Tipon, and J. B. Whelan. 1978. Evaluation of a bear hunting
survey and its implications on black bear management in Virginia. Proceedings

of the Annual Conference of Southeastern Association of Fish and Wildlife
Agencies 32: 202—207.

Duda, M. D., and K. L. Brown. 2001. Texas landowners’ attitudes toward wildlife,

conservation, and outdoor recreation. Transactions of the 66th North American
Wildlife and Natural Resources Conference 66: 96-109.

Duda, M. D. 1986. Wildlife management-the human element. Florida Wildlife.
September/October.

Dytham, C. 2003. Choosing and using statistics. Second edition. Blackwell Publishing,
Malden, Massachusetts, USA.

Eagle, T. C., and M. R. Pelton. 1980. Seasonal nutrition of black bears in the Great

Smoky Mountains National Park. International Conference on Bear Research and
Management 5: 94—101.

Eastridge, R. 2000. Experimental repatriation of black bears to the Big South Fork Area

of Kentucky and Tennessee. Thesis, The University of Tennessee, Knoxville,
Tennessee, USA.

Elkie, P., R. Rempel, and A. Carr. 1999. Patch Analyst User’s Manual. Ontario Ministry
of Natural Resources Northwest Science & Technology. Thunder Bay, Ontario,
Canada. TM-002.

Enck, J. W. and T. L. Brown. 2002. New Yorkers' attitudes toward restoring wolves to
the Adirondack Park. Wildlife Society Bulletin 30: 16—28.

Enck, J. W., and T. L. Brown. 2000. Preliminary assessment of social feasibility for
reintroducing gray wolves to the Adirondack Park in northern New York. HDRU
Series No. 00—3. Cornell University, Ithaca, NY.

Epps, C. W. 1997. Habitat suitability for black bear (Ursus americanus) in the Neches
Bottom and Jack Gore Baygall Units of the Big Thicket National Preserve.

Senior Honors Thesis. Rice University, Houston, Texas, USA.

F ahri g, L., and G. Merriam. 1985. Habitat patch connectivity and population survival.
Ecology 66: 1762-1768.

Feis, M. L., D. D. Martin, and G. T. Blank, Jr. 1987. Movements and rates of return of

252

translocated black bears in Virginia. International Conference on Bear Research
and Management 7: 369—372.

Fisher, J .T., and L. Wilkinson. 2005. The response of mammals to forest fire and timber
harvest in the North American boreal forest. Mammal Review 35: 51—81.

Foster, M. L., and S. R. Humphrey. 1995. Use of highway underpasses by Florida
panthers and other wildlife. Wildlife Society Bulletin 23: 95-100.

Fox, J., P. Yonzon, and N. Podger. 1996. Mapping conflicts between biodiversity and
human needs in Langtang National Park, Nepal. Conservation Biology 10: 562—
569.

Garshelis, D. L., and M. R. Pelton. 1980. Activity of black bears in the Great Smoky
Mountains National Park. Journal of Mammalogy 61: 8—19.

Garshelis, D. L., and M. R. Pelton. 1981. Movement of black bears in the Great Smoky
Mountains National Park. Journal of Wildlife Management 45: 912-925.

Garner, N. P. 1996. Suitability of habitats in east Texas for black bears. Texas Parks
and Wildlife Department. Federal Aid Grant No. W—125—R-7.

Gilpin, M. E., and J. M. Diamond. 1980. Subdivision of nature reserves and the
maintenance of species diversity. Nature 285: 567-568.

Gunter, P. 1971. The Big Thicket: a Challenge for Conservation. The Jenkins Book
Publishing Company, Inc., Austin, Texas, USA.

Gunter, P. A. 1993. The Big Thicket: an Ecological Reevaluation. University of North
Texas Press, Denton, Texas, USA.

Gurd, D. B., and T. D. Nudds. 1999. Insular biogeography of mammals in Canadian
parks: a re-analysis. Journal of Biogeography 26: 973-982.

Gurd, D. B., T. D. Nudds, and D. H. Rivard. 2001. Conservation of mammals in eastern
North American wildlife reserves: how small is too small? Conservation Biology
15: 1355—1363.

Haddad, N. M. 1999. Corridor use predicted from behaviors at habitat boundaries. The
American Naturalist 153: 215-227.

Haefner, J. W. 1996. Modeling biological systems: principles and applications.
Chapman and Hall, Inc., New York, New York, USA.

Hagar, J ., S. Howlin, and L. Ganio. 2004. Short-term response of songbirds to

253

experimental thinning of Douglas-fir forests in the Oregon Cascades. Forest
ecology and management 199: 333—347.

Hall, E. R. 1981. The Mammals of North America. John Wiley and Sons, New York,
New York, USA.

Hamilton, D. 1999. Controversy in times of plenty. Missouri Conservationist 60: 17-23.

Hamilton, R. J. 197 8. Ecology of the black bear in southeast North Carolina. Thesis,
University of Georgia, Athens, Georgia, USA.

Hamilton, R. J ., & R. L. Marchinton. I980. Denning and related activities of black bears

in the coastal plain of North Carolina. Bear Biology Association Conference
Series 3: 121—126.

Hamin, E. M. 2001. The US National Park Service’s partnership parks: collaborative
responses to middle landscapes. Land Use Policy 18: 123-125.

Hansen, A. J ., and J. J. Rotella. 2002. Biophysical factors, land use, and species viability
in and around nature reserves. Conservation Biology 16: 1112-1122.

Harcombe, P. A., and G. Callaway. 1997. Management assessment of the water corridor
units of the Big Thicket National Preserve. Final Report; National Park Service,
Beaumont, Texas, USA.

Harcombe, P. A., and P. L. Marks. 1977. Understory structure of a mesic forest in
southeast Texas. Ecology 58: 1144-1151.

Harcourt, A. H., S. A. Parks, and R. Woodroffe. 2001. Human density as an influence
on species/area relationships: double jeopardy for small African reserves?
Biodiversity and Conservation 10: 1011—1026.

Hastings, B. C., and W. E. Hammitt. 1986. Preferences of visitors for wildlife species.
Pages 89—93 in Kulhavy, D. L., and R. N. Conner, editors. Wilderness and
natural areas in the eastern United States: a management challenge. Center for
Applied Studies, School of Forestry, Stephen F. Austin State University,
Nacogdoches, Texas, USA.

Hellgren, E. C., and M. R. Vaughan. 1989a. Demographic analysis of a black bear
population in the Great Dismal Swamp. The Journal of Wildlife Management 53:
969—977.

Hellgren, E. C., and M. R. Vaughan. 1989b. Denning ecology of black bears in a
southeastern wetland. The Journal of Wildlife Management 53: 347—353.

Hellgren, E. C., and M. R. Vaughan. 1990. Range dynamics of black bears in Great

254

Dismal Swamp, Virginia-North Carolina. Proceedings of the Annual Conference
of Southeastern Fish and Wildlife Agencies 44: 268-278.

Hellgren, E. C., M. R. Vaughan, and D. F. Stauffer. 1991. Macrohabitat use by black
bears in a southeastern wetland. Journal of Wildlife Management 55: 442-448.

Higgs, A. J. 1981. Island biogeography theory and nature reserve design. Journal of
Biogeography 8: 117-124.

Hillman, L. L., and Yow, D. L. 1986. Timber management for black bear. Proceedings
from the 8‘h Eastern Black Bear Workshop on Research and Management.
Williamsburg, Virginia, USA.

.Holling, C. S., and G. K. Meffe. 1996. Command and control and the pathology of
natural resource management. Conservation Biology 10: 328—337.

Johnson, K. G., and M. R. Pelton. 1981. Selection and availability of dens for black
bears in Tennessee. Journal of Wildlife Management 45: 111—119.

Joly, D. O., C. A. Ribic, J. A. Langenberg, K. Beheler, C. A. Batha, B. J. Dhuey, R. E.
Rolley, G. Bartlelt, T. R. Van Deelen, and M. D. Samual. 2003. Chronic wasting

disease in free-ranging Wisconsin white-tailed deer. Emerging Infectious Disease
9: 599—601.

Jones, M. D., G. S. Warburton, and M. R. Pelton. 1998. Models for predicting occupied
black bear habitat in coastal North Carolina. Ursus 10: 203—207.

J onker, S. A., J. A. Parkhurst, R. Field, and T. K. Fuller. 1998. Black bear depredation
on agricultural commodities in Massachusetts. Wildlife Society Bulletin 26: 318—
324.

Karl, J. W., P. J. Heglund, E. O. Garton, J. M. Scott, N. M. Wright, and R. L. Hutto.
2000. Sensitivity of species habitat-relationship model performance to factors of

scale. Ecological Applications 10: 1690-1705.

Kellert, S. R. 1994. Public attitudes toward bears and their conservation. International
Conference on Bear Research and Management 9: 43—50.

Kellert, S. R. 1991. Public views of wolf restoration in Michigan. Transactions of the
North American Wildlife and Natural Resources Conference 56: 152—1 61.

Kellert, S. R. 1985a. Social and perceptual factors in endangered species management.
Journal of Wildlife Management 49: 528—536.

Kellert, S. R. 1985b. Public perceptions of predators, particularly the wolf and coyote.
Biological Conservation 3 1 : 167—189.

255

Kellert, S. R. 1985c. The public and the timber wolf in Minnesota. Transactions of the
North American Wildlife and Natural Resources Conference 51: 193—200.

Kellert, S. R., M. Black, C. R. Rush, and A. J. Bath. 1996. Human culture and large
carnivore conservation in North America. Conservation Biology 10: 977—990.

Kennedy, M. L. 1989. Systematic assessment of the black bear, Ursus americanus
luteolus. Special report to US. Fish and Wildlife Service. Jackson Field Office,
Jackson, Mississippi, USA.

Kie, J. G., R. T. Bowyer, M. C. Nicholson, B. B. Boroski, and E. R. Loft. 2002.
Landscape heterogeneity at differing scales: effects on spatial distribution of mule
deer. Ecology 83: 530—544.

Kleiman, D. G. 1989. Reintroduction of captive mammals for conservation. Bioscience
39: 152—161.

Knight, R. L. 1999. Private lands: the neglected geography. Conservation Biology 13:
223—224.

Knox. E. G. 2005. Childhood cancers and atmospheric carcinogens. Journal of
Epidemiology and Community Health 59: 101—105.

Koch, DB. 1994. Biopolitical management of mountain lions, Tule elk, and black bears
in California. International Conference on Bear Research and Management 9:
561—566.

Koval, M. H., and A. G. Mertig. 2004. Attitudes of the Michigan public and wildlife
agency personnel toward lethal wildlife management. Wildlife Society Bulletin
32: 23 2—243.

Kulhavy, D. L., and R. N. Conner, editors. 1986. Wilderness and natural areas in the
eastern United States: a management challenge. Center for Applied Studies,
School of Forestry, Stephen F. Austin State University, Nacogdoches, Texas,
USA.

Kulldorff, M., W. Athas, E. F euer, B. Miller, and C. Key. 1998. Evaluating cluster
alarms: a space-time scan statistic for brain cancer in Los Alamos. American

Journal of Public Health 88: 1377—1380.

Kulldorff, M., and N. Nagarwalla. 1995. Spatial disease clusters: detection and
inference. Statistics and Medicine 14: 799—810.

Kulldorff, M. 1997. A spatial scan statistic. Communications in Statistics: Theory and
Methods 26: 1481—1496.

256

Kulldorff, M., E. J. Feuer, B. A. Miller, and L. S. Freedman. 1997. Breast cancer
clusters in the Northeast United States: a geographic analysis. American Journal
of Epidemiology 146: 161—170.

Landers, J. L., R. J. Hamilton, A. S. Johnson, and R. L. Marchinton. 1979. Food and
habitat of black bears in southeastern North Carolina. The Journal of Wildlife
Management 43: 143—153.

Lariviére, S. 2001. Ursus americanus. Mammalian Species, No. 647. American
Society of Mammalogists.

Larkin, J. L., D. S. Maehr, T. S. Hoctor, M. A. Orlando, and K. Whitney. 2004.
Landscape linkages and conservation planning for the black bear in west-central
Florida. Animal Conservation 7: 23—34.

Lauber, T. B., and B. A. Knuth. 1996. Citizens’ and agency staff members’ evaluation
of decision-making procedures: a case study of the New York State moose
reintroduction issue. HDRU Series No. 96-1. Cornell University, Ithaca, New
York, USA.

Leader-Williams, N., J. Harrison, and M. J. B. Green. 1990. Designing protected areas
to conserve natural resources. Science Progress Oxford 74: 189—204.

Lee, M. E., and R. Miller. 2003. Managing elk in the wildland-urban interface: attitudes
of Flagstaff, Arizona residents. Wildlife Society Bulletin 31: 185:191.

Linnell, J. D. C., J. E. Swenson, R. Andersen, and B. Barnes. 2000. How vulnerable are
denning bears to disturbance. Wildlife Society Bulletin 28: 400-413.

Liu. J ., G. C. Daily, P. R. Ehrlich, and G. W. Luck. 2003. Effects of household
dynamics on resource consumption and biodiversity. Nature 421: 530-533.

Liu, J ., and W. W. Taylor. 2002. Integrating landscape ecology into natural resource
management. Cambridge University Press, Cambridge, England.

Liu, J ., M. Linderman, Z. Ouyang, L. An, J. Yang, and H. Zhang. 2001. Ecological
degradation in protected areas: the case of Wolong Nature Reserve for Giant
Pandas. Science 292: 98—101.

Liu. J. Z. Ouyang, W. Taylor, R. Groop, Y. Tan, and H. Zhang. 1999. A framework for
evaluating effects of human factors on wildlife habitat: the case of giant pandas.

Conservation Biology 13: 1360—1370.

Liu, J. 1993. ECOLECON: An ECOLogical-ECONomic model for species conservation
in complex forest landscapes. Ecological Modelling 70: 63—87.

257

Liu, J ., F. W. Cubbage, and H. R. Pulliam. 1994. Ecological and economic effects of
forest landscape structure and rotation length: simulation studies using
ECOLECON. Ecological Economics 10: 249—263.

Liu, J ., J. B. Dunning, and H. R. Pulliam. 1995. Potential effects of a forest management
plan on Bachman’s sparrows (A imophila aestivalis): linking a spatially explicit
model with GIS. Conservation Biology 9: 62—75.

Locker, C. A., and D. J. Decker. 1995. Colorado black bear hunting referendum: what
was behind the vote? Wildlife Society Bulletin 23: 370—376.

Lohr, C., W. B. Ballard, and A. Bath. 1996. Attitudes toward gray wolf reintroduction to
New Brunswick. Wildlife Society Bulletin 24: 414—420.

Maehr, D. S., and J. R. Brady. 1984. Food habits of Florida black bears. Journal of
Wildlife Management 48: 230—235.

Maehr, D. S., J. S. Smith, M. W. Cunningham, M. E. Barnwell, J. L. Larkin, and M. A.
Orlando. 2003. Spatial characteristics of an isolated Florida black bear
population. Southeastern Naturalist 2:443—446.

Manfredo M. J ., H.C. Zinn, L. Sikorowski, and J. Jones. 1998. Public acceptance of
mountain lion management: a case study of Denver, Colorado, and nearby
foothills areas. Wildlife Society Bulletin 26: 964-970.

Marchinton, F. B. 1995. Movement ecology of black bears in fragmented bottomland
hardwood habitat in Louisiana. Thesis, The University of Tennessee, Knoxville,
Tennessee, USA.

Margai, F ., and N. Henry. 2003. A community-based assessment of learning disabilities
using environmental and contextual risk factors. Social Science & Medicine 56:
1073—1085.

Marks, P. L., and P. A. Harcombe. 1981. Forest vegetation of the Big Thicket, southeast
Texas. Ecological Monographs 51:287-305.

Mather, T. N., M. C. Nicholson, R. Hu, and N. J. Miller. 1996. Entomological correlates
of Babesia microti prevalence in an area where Ixodes scapularis (Acari:
Ixodidae) is endemic. Journal of Medical Entomology 33: 866—870.

Mazerolle, M. J ., and M. A. Villard. 1999. Patch characteristics and landscape context
as predictors of species presence and abundance: a review. Ecoscience 6:117-

124.

McDonald, A. A., J. Liu, H. Prince, and K. Kress. 2001. A socio-economic-ecological

258

simulation model of land acquisition to expand a national wildlife refuge.
Ecological Modelling 140: 99—110.

McLellan, B. N., and F. W. Hovey. 2001. Habitats selected by grizzly bears in a
multiple use landscape. Journal of Wildlife Management 65: 92-99.

Means, J. E., and S. E. Greene. 1987. Reserve designs in a landscape context: another
slant in the continuing debate. Park Science 8: 21.

Meffe, G. K., and C. R. Carroll. 1997. Principles of Conservation Biology. Second
edition. Sinauer Associates, Inc. Publishers, Sunderland, Massachusetts, USA.

Merrill, T., D. J. Mattson, R. G. Wright, and H. B. Quigley. 1999. Defining landscapes
suitable for restoration of grizzly bears Ursus arctos in Idaho. Biological
Conservation 87: 231—248.

Mertig, A. G. 2004. Attitudes about wolves in Michigan, 2002. Report to the Michigan
Department of Natural Resources, Wildlife Division. Lansing, Michigan, USA.

Messmer, T. A., M. W. Brunson, D. Reiter, and D. G. Hewitt. 1999. United States
public attitudes regarding predators and their management to enhance avian
recruitment. Wildlife Society Bulletin 27: 75-85.

Miller, J. R., M. G. Turner, E. A. H. Smithwick, C. L. Dent, E. H. Stanley. 2004. Spatial
extrapolation: the science of predicting ecological patterns and processes.
Bioscience 54: 310-320.

Miller, M. A., I. A. Gardner, C. Kreuder, D. M. Paradies, K. R. Worcester, D. A. Jessup,
E. Dodd, M. D. Harris, J. A. Ames, A. E. Pankham, and P. A. Conrad. 2002.
Coastal freshwater runoff is a risk factor for Toxoplasma gondii infection of

southern sea otters (Enhydra lutris nereis). International Journal of Parasitology
32: 997—1006.

Moilanen, A., and I. Hanski. 2001. On the use of connectivity measures in spatial
ecology. Oikos 95: 147-151.

Morzillo, A. T., J. Liu, and A. G. Mertig. 2005. Attitudes toward and opinions about
black bears in East Texas. Technical Report, prepared for the Texas Parks and
Wildlife Department.

Mostashari, F., M. Kulldorff, J. J. Hartman, J. R. Miller, and V. Kulasekera. 2003. Dead
bird clusters as an early warning system for West Nile Virus activity. Emerging

Infectious Diseases 9: 641—646.

Murray, D. L., J. D. Roth, E. Ellsworth, A. J. Wirsing, and T. D. Steury. 2002.

259

Estimating low-density snowshoe hare populations using fecal pellet counts.
Canadian Journal of Zoology 80: 771-781.

National Parks Conservation Association (NPCA). 2004. NPCA Press Releases. Big
Thicket Named to List of Endangered National Parks. Accessed: May 17, 2005.
<http://www.npca.org/media center/bigthicket.asp>

Naughton-Treves, L. 1997. Farming the forest edge: vulnerable places and people
around Kibale National Park, Uganda. Geographical Review 87: 27—46.

Neal, W. A. 1990. Proposed threatened status for the Louisiana black bear. Federal
Register, 21 June 1990. 50 CFR Part 17 55(120): 25341—25345.

Newmark, W. D. 1985. Legal and biotic boundaries of western North American national
parks: a problem of congruence. Biological Conservation 33: 197-208.

Newmark, W. D. 1995. Extinction of mammal populations in western North American
national parks. Conservation Biology 9: 512-526.

Newmark, W. D. 1996. Insularization of Tanzanian parks and the local extinction of
large mammals. Conservation Biology 10: 1549—1556.

Noble, 1. R., and R. Dirzo. 1997. Forests as hurnan-dominated ecosystems. Science
277: 522—525.

Norris, D. R., M. T. Theberge, and J. B. Theberge. 2002. Forest composition around
wolf (Canis lupus) dens in eastern Algonquin Provincial Park, Ontario. Canadian
Journal of Zoology 80: 866-872.

Norton, D. A. 2002. Conservation biology and private land: shifting the focus.
Conservation Biology 14: 1221—1223.

Noss, R. F., C. Carroll, K. Vance-Borland, and G. Wuerthner. 2002. A multicriteria
assessment of the irreplaceability and vulnerability of sites in the Greater
Yellowstone Ecosystem. Conservation Biology 16: 895—908.

Nyland, P. D. 1995. Black bear habitat relationships in coastal Louisiana. Thesis,
Louisiana State University, Baton Rouge, Louisiana, USA.

O’Connell, M. A., and R. F. Noss. 1992. Private land management for biodiversity
conservation. Environmental Management 16: 435—450.

Oli, M. K., H. A. Jacobson, and B. D. Leopold. 1997. Denning ecology of black bears in

the White River National Wildlife Refuge, Arkansas. Journal of Wildlife
Management 61: 700—706.

260

Osborne, P. E., J. C. Alonso, and R. G. Bryant. 2001. Modelling landscape-scale habitat
use using GIS and remote sensing: a case study with great bustards. Journal of
Applied Ecology 38: 458-471.

Pacelle, W. 1998. Foraging a new wildlife management paradigm: Integrating animal
protection values. Human Dimensions of Wildlife 3: 42-50.

Parks, S. A., and A. H. Harcourt. 2002. Reserve size, local human density, and

mammalian extinctions in US protected areas. Conservation Biology 16: 800—
808.

Pate, J. M. J. Manfredo, A. (1. Bright, and G. Tischbein. 1996. Coloradans’ attitudes
toward reintroducing the gray wolf in Colorado. Wildlife Society Bulletin 24:
421—428.

Peacock, H. 1994. Nature lover’s guide to the Big Thicket. Texas A&M University
Press, College Station, Texas, USA.

Pelton, M. R. 1986. Habitat needs of black bears in the east. Pages 49—53 in D. L.
Kulhavy and R. N. Conner, editors. Wilderness and Natural Areas in the Eastern
United States: a management challenge. Center for Applied Studies, School of
Forestry, Stephen F. Austin State University, Nacogdoches, Texas, USA.

Pelton, M. R. 1982. Black bear. Pages 504—514 in J. A. Chapman and G. A. Feldhamer,
editors. Wild mammals of North America. First edition. John Hopkins
University Press, Baltimore, Maryland, USA.

Pelton, M. R., C. D. Scott, and G. M. Burghardt. 1976. Attitudes and opinions of
persons experiencing property damage and/or injury by black bears in the Great
Smoky Mountains National Park. International Conference on Bear Research and
Management 3: 157—167.

Peterson, D. L., and V. T. Parker, editors. 1998. Ecological scale. Columbia University
Press, New York, New York, USA.

Peyton, B., P. Bull, T. Reis, and L. Visser. 2001. Public views on bear and bear
management in the lower peninsula of Michigan. Report to the Wildlife Division,

Michigan Department of Natural Resources, Lansing, Michigan, USA.

Pickett. S. T. A., and J. N. Thompson. 1978. Patch dynamics and the design of nature
reserves. Biological Conservation 13:27-37.

Pimm, S. L., and J. H. Lawton. I998. Ecology-planning for biodiversity. Science
279: 2068—2069.

Pressey, R. L., G. L. Whish, T. W. Barrett, and M. E. Watts. 2002. Effectiveness of

261

protected areas in north—eastem New South Wales: recent trends in six measures.
Biological Conservation 106: 57—69.

Pressey, R. L., and S. L. Tully. 1994. The cost of ad hoc reservation: a case study in
western New South Wales. Australian Journal of Ecology 19: 375—3 84.

Pritchard, G. T., and C. T. Robbins. 1990. Digestive and metabolic efficiencies of
grizzly and black bears. Canadian Journal of Zoology 68: 1645-1651.

Purdy, K. G., and D. J. Decker. 1989. Obtaining wildlife values information for
management: the wildlife attitudes and values scale (WAVS). HDRU Series
Number 89—2. Cornell University, Ithaca, New York, USA.

Ratti, J. T., M. Weinstein, J. M. Scott, P. A. Wiseman, An. M. Gillesberg, C. A. Miller,
M. M. Szepanski, and L. K. Svancara. 2004. Feasibility of wolf reintroduction to
Olympic Peninsula, Washington. Northwest Science 78: 1-76.

Reading, R. P., and T.W. Clark. 1996. Carnivore reintroductions: and interdisciplinary
examination. Pages 296—336 in J. L. Gittleman, editor. Carnivore behavior,
ecology, and evolution. Volume 2. Cornell University Press, Ithaca, New York,
USA.

Reading, R. P., and S. R. Kellert. 1993. Attitudes toward a proposed reintroduction of
black-footed ferrets (Mustela nigripes). Conservation Biology 7: 569—580.

Reynolds, D. G., and J. J. Beecham. 1980. Home range activities and reproduction of
black bears in west-central Idaho. Bear Biologist Association Conference Series
3: 181-191.

Riffell, S. K., B. E. Keas, and T. M. Burton. 2003. Birds in North American Great Lakes
coastal wet meadows: is landscape context important? Landscape Ecology 18: 95-
111.

Riley, S. J ., K. Aune, R. D. Mace, and M. J. Madel. 1994. Translocation of nuisance
grizzly bears in northwestern Montana. International Conference on Bear
Research and Management 9: 567—573.

Riley, S. J ., and D. J. Decker. 2000a. Risk perception as a factor in wildlife stakeholder
acceptance capacity for cougars in Montana. Human Dimensions of Wildlife 5:
50—62.

Riley, S. J ., and D. J. Decker. 2000b. Wildlife stakeholder acceptance capacity for
cougars in Montana. Wildlife Society Bulletin 28: 931—939. ch 6

Riley, S. J ., D. J. Decker, L. H. Carpenter, J. F. Organ, W. F. Siemer, G. F ., Mattfeld, and

262

 

G. Parsons. 2002. The essence of wildlife management. Wildlife Society
Bulletin 30: 585—593.

Rivard, D. H., J. Poitevin, D. Plasse, M. Carleton, and D. J. Currie. 2000. Changing
species richness and composition in Canadian National Parks. Conservation
Biology 14: 1099—1109.

Rogers, M. J. 1973. Movements and reproductive success of black bears introduced into
Arkansas. Proceedings of the Southeastern Association of Fish and Wildlife
Agencies 27: 307—308.

Rogers, L. L. 1986. Effects of translocation distance on frequency of return by adult
black bears. Wildlife Society Bulletin 14: 76—80.

Rogers, L. L. 1987. Effects of food supply and kinship on social behavior, movements,
and population grth of black bears in northeastern Minnesota. Wildlife
Monographs 97: 1—72.

Rogers, LL, and AW. Allen. 1987. Habitat suitability index models: black bear, upper
Great Lakes region. US. Fish and Wildlife Service. Biological Report
82(10. 144).

Rudis, VA. 1986. Emerging patterns in the distribution of roadless forested areas in the
Midsouth. Pages 265—270 in Kulhavy, D.L., and RN. Conner, editors.
Wilderness and natural areas in the eastern United States: a management
challenge. Center for Applied Studies, School of Forestry, Stephen F. Austin
State University, Nacogdoches, Texas, USA.

Rudis, V. A., and J. B. Tansey. 1995. Regional assessment of remote forest and black
bear habitat from forest remote surveys. Journal of Wildlife Management 59:
170—180.

Sauer, P. R., S. L. Free, and S. D. Browne. 1969. Movement of tagged black bears in the
Adirondacks. NewYork Fish and Game Journal 16: 205—223.

Schamberger, M. A., H. Farmer, and J. W. Terrell. 1982. Habitat suitability index
models: introduction. US. Fish and Wildlife Service. F WS/OBS—82/ 10.

Schoenecker, K. A. and W. W. Shaw. 1997. Attitudes toward a proposed reintroduction
of Mexican gray wolves in Arizona. Human Dimensions of Wildlife 2: 42—55.

Schonewald-Cox, C. 1988. Boundaries and the protection of nature reserves.
Bioscience 38: 480-486.

Schonewald-Cox, C. M., and J. W. Bayless. 1986. The boundary model: a geographic

263

 

analysis of design and conservation of nature reserves. Biological Conservation
38: 305-322.

Schonewald-Cox, C., M. Buechner, R. Sauvajot, and B. A. Wilcox. 1992. Cross-
boundary management between national parks and surrounding lands: a review
and discussion. Environmental Management 16: 273-282.

Schullery, P. 1986. The bears of Yellowstone. Robert Reinhart, Inc. Publishers,
Boulder, Colorado.

Schumaker, N. H. 1996. Using landscape indices to predict habitat connectivity.
Ecology 77: 1210-1225.

Sellars, R. W. 1997. Preserving Nature in the National Parks: a History. Yale
University Press, New Haven.

Schooley, R. L., C. R. McLaughlin, W. B. Krohn, G. J. Matula, Jr. 1994. Spatiotemporal
patterns of macrohabitat use by female bears during fall. International
Conference on Bear Research and Management 9: 339—348.

Shafer, C. L. 1999. US national park buffer zones: historical, scientific, social, and legal
aspects. Environmental Management 23: 49-73.

Shriver, W. G., T. P. Hodgman, J. P. Gibbs, and P. D. Vickery. 2004. Landscape context

influences salt marsh bird diversity and area requirements in New England.
Biological Conservation 119: 545-553.

Shropshire, C. C. 1996. History, status, and habitat components of black bears in
Mississippi. Dissertation, Mississippi State University, Mississippi State,
Mississippi, USA.

Shull, S. D. 1994. Management of nuisance black bears (Ursus americanus) in the
Interior Highlands of Arkansas. Thesis, University of Arkansas, F ayetteville,
USA.

Sillero-Zubiri, C., and M. K. Laurenson. 2001. Interactions between carnivores and
local communities: conflict or co-existence. Pages 282—312 in J. Gittleman, S. M.
Funk, D. Macdonald, and R. K. Wayne, editors. Carnivore conservation.
Cambridge University Press, Cambridge, England.

Simberloff, D., and J. Cox. 1987. Consequences and costs of conservation corridors.
Conservation Biology 1: 63-71.

Sitati, N. W., M. J. Walpole, R. J. Smith, and N. Leader-Williams. 2003. Predicting

spatial aspects of human-elephant conflict. Journal of Applied Ecology 40: 667—
677.

264

 

Smith, T. R. 1985. Ecology of black bears in bottomland hardwood forest in Arkansas.
Dissertation, The University of Tennessee, Knoxville, Tennessee, USA.

Smith, K. G., and J. D. Clark. 1994. Black bears in Arkansas: characteristics of a
successful translocation. Journal of Mammalogy 75: 309—320.

Smith, K. L., V. DeVos, H. Bryden, L. B. Price, M. E. Hugh-Jones, and P. Keim. 2000.
Bacillus anthracis diversity in Kruger National Park. Journal of Clinical
Microbiology 38: 3780—3784.

Sokal, R. R., and F. J. Rohlf. 1995. Biometry. Third edition. W. H. Freeman and
Company, New York, New York, USA.

Soule, M. E. 1991. Conservation: tactics for a constant crisis. Science 253: 744—750.

Soule, M. E., and MA. Sanjayan. 1998. Ecology-conservation targets: do they help?
Science 279: 2060—2061.

Soulé. M. E., and D. Simberloff. 1986. What do genetics and ecology tell us about the
design of nature reserves? Biological Conservation 35: 19—40.

South, A. B., S. P. Rushton, D. W. Macdonald, and R. Fuller. 2001. Reintroduction of
the European beaver (Castor fiber) to Norfolk, UK: a preliminary modeling
analysis. Journal of Zoology London 254: 473—479.

Stern, S. J. 1998. Field studies of large mobile organisms: scale, movement, and habitat
utilization. Pages 289-307 in Peterson, D. L., and V. T. Parker, editors. 1998.
Ecological scale. Columbia University Press, New York, New York, USA.

Steury, T. D., and D. L. Murray. 2004. Modeling the reintroduction of lynx to the
southern portion of its range. Biological Conservation 117: 127—141.

Stinson, P. H., and R. M. Pace, III. 1995. A survey of the tunica Hills region for black
bear: Phase II completion report. Louisiana State University, Baton Rouge.
Report to US. Fish and Wildlife Service, Jackson, Mississippi, USA.

Stinson, P. H. 1996. Black bears in the Tunica Hills: present status and habitat
suitability. Thesis, Louisiana State University, Baton Rouge, Louisiana, USA.

Stiver, W. H. 1991. Population dynamics and movements of problem black bears in
Great Smoky Mountains National Park. Thesis, The University of Tennessee,
Knoxville, Tennessee, USA.

Stubblefield, C. H. 1993. Food habits of black bears in the San Gabriel Mountains of
southern California. The Southwestern Naturalist 38: 290—293.

265

 

Teel, T. L., R. S. Krannich, and R. H. Schmidt. 2002. Utah stakeholders’ attitudes
toward selected cougar and black bear management practices. Wildlife Society
Bulletin 30: 2-15.

Texas Parks and Wildlife Department (TPWD). 2005. East Texas black bear
conservation and management plan. Austin, Texas, USA. PWD PL W7000-1046
(6/05).

Tischendorf, L., and L. Fahrig. 2000a. How should we measure landscape connectivity?
Landscape Ecology 15: 633-641.

Tischendorf, L., and L. Fahrig. 2000b. On the usage and measurement of landscape
connectivity. Oikos 90: 7-19.

Texas Parks and Wildlife Department (TPWD). 2005. East Texas black bear
conservation and management plan. Austin, Texas, USA. PWD PL W7000-1046
(6/05).

Truett, J. C., and D. W. Lay. 1984. Land of bears and honey: 3 natural history of East
Texas. University of Texas Press, Austin, Texas, USA.

Turner, M. G., R. H. Gardner, and R. V. O’Neill. 2001. Landscape ecology in theory
and practice. Springer-Verlag New York, Inc., New York, New York, USA.

Turner, M. G., R. H. Gardner, and R. V. O’Neill. 1995. Ecological dynamics at broad
scales. BioScience Supplementary, $29—$35.

Turner, M. G., R. V. O’Neill, R. H. Gardner, and B. T. Milne. 1989. Effects of changing
spatial scale on the analysis of landscape pattern. Landscape Ecology 3: 153-162.

US. Fish and Wildlife Service. 1981. Standards for the development of habitat
suitability index models for use in the habitat evaluation procedures. Division of
Ecological Services ESM 103.

van Manen, F. T. 1991. A feasibility study for the potential reintroduction of black bears
into the Big South Fork Area of Kentucky and Tennessee. Tennessee Wildlife
Resources Agency Technical Report No. 9103, Morristown, Tennessee, USA.

van Manen, F. T., and M. R. Pelton. 1997. A GIS model to predict black bear habitat
use. Journal of Forestry 95: 6—12.

Van Why, K. R. 2003. Feasibility of restoring the Louisiana black bear (Ursus

americanus luteolus) to portions of their historic range. Thesis, Louisiana State
University, Baton Rouge, Louisiana, USA.

266

 

Van Why, K. R., and M. J. Chamberlain. 2003. Hunter knowledge and opinion of
Louisiana black bear restoration. Proceedings of the Annual Conference of
Southeastern Fish and Wildlife Agencies 57: 299—307.

Vitousek, P. M., H. A. Mooney, J. Lubchnco, and J. M. Melillo. 1997. Human
domination of Earth’s ecosystems. Science 277: 494—499.

Voogd, H. 1983. Multicriteria evaluation for urban and regional planning. Pion
Limited, London, UK.

Wagner, R. O. 1990. Movement patterns of black bears in south central Louisiana.
Thesis, Louisiana State University, Baton Rouge, Louisiana, USA.

Weaver, K. M., D. K. Tabberer, L. U. Moore Jr., G. A. Chandler, J. C. Posey, and M. R.
Pelton. 1990. Bottomland hardwood forest management for black bears in

Louisiana. Proceedings of the Annual Conference of Southeastern Association of
Fish and Wildlife Agencies 44: 342—350.

Weaver, K. M. 1992. The ecology and management of black bears in the Tensas River
Basin of Louisiana. Dissertation, The University of Tennessee, Knoxville,
Tennessee, USA.

White, T. H., Jr., J. L. Bowman, B. D. Leopold, H. A. Jacobson, W. P. Smith, and F. J.
Vilella. 2000. Influence of Mississippi alluvial valley rivers on black bear
movements and dispersal: implications for Louisiana black bear recovery.
Biological Conservation 95: 323-331.

Wiegand, T., K. A. Moloney, J. Naves, and F. Knauer. 1999. Finding the missing link
between landscape structure and population dynamics: a spatially explicit
perspective. The American Naturalist 154: 605-627.

With, K. A., R. H. Gardner, and M. G. Turner. 1997. Landscape connectivity and
population distributions in heterogeneous environments. Oikos 78: 151-169.

Wooding, J. B., and T. S. Hardisky. 1992. Home range, habitat use, and mortality of
black bears in north-central Florida. Journal of Mammalogy 73: 895—898.

Wooding, J. B., J. A. Cox, and M. R. Pelton. 1994a. Distribution of black bears on the
southeastern coastal plain. Proceedings of the Annual Conference of the

Southeastern Association of Fish and Wildlife Management Agencies 48: 270—
275. ch]

Wooding, J. B., T. S. Hardisky, J. J. Clarr, and P. Schullery. 1994b. Home range, habitat

use, and mortality of black bears in north-central Florida. Proceedings of the 9th
lntemational Conference on Bear Research and Management 9: 349-3 56. ch.2

267

 

 

Woodroffe, R. 2001a. Strategies for carnivore conservation: lessons from contemporary
extinctions. Pages 61—92 in J. Gittleman, S. M. Funk, D. Macdonald, and R. K.
Wayne, editors. Carnivore conservation. Cambridge University Press,
Cambridge, England.

Woodroffe, R. 2001b. Predators and people: using human densities to interpret declines
of large carnivores. Animal Conservation 3: 165-173.

Woodroffe, R., and J. R. Ginsberg. 1998. Edge effects and the extinction of populations
inside protected areas. Science 280: 2126—2128.

Wright, R. G. 1999. Wildlife management in the national parks: questions in search of
answers. Ecological Applications 9: 30—36.

Xie, J. L., H. R. Hill, S. R. Winterstein, H. Campa, R. V. Doepker, T. R. Van Deelen, and

J. G. Liu. 1999. White-tailed deer management options model (DeerMOM):
design, quantification, and application. Ecological Modelling 124: 121—130.

268

 

 

 

 

 

   

lil‘lj‘fljijl‘lflijflijljjjl

,'........-n---—-——--